(navigation image)
Home American Libraries | Canadian Libraries | Universal Library | Community Texts | Project Gutenberg | Children's Library | Biodiversity Heritage Library | Additional Collections
Search: Advanced Search
Anonymous User (login or join us)
Upload
See other formats

Full text of "Adsorption mechanism(s) of poly(ethylene oxide) on oxide and silicate surfaces"

ADSORPTION MECHANISM(S) OF POLYETHYLENE OXIDE) ON 
OXIDE AND SILICATE SURFACES 



By 

SHARAD MATHUR 



A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE 

UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE 

REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY 



UNIVERSITY OF FLORIDA 
1996 



ACKNOWLEDGMENTS 

I would like to express my deepest gratitude to Dr. B.M. Moudgil, my major 
advisor, for his invaluable guidance, assistance and encouragement through this 
investigation. 

My sincere thanks go to Dr. DO. Shah, Dr. E.D. Whitney, Dr. CD. Batich, 
Dr. H. El-Shall and Dr, R.K. Singh for serving on my supervisory committee. 

To all my friends- Dr. S.Behl, Dr. R.Damodaran, N.Kulkami, T.S. Prakash, 
S.Zhu, R.Kalyanraman, J. Adler and many more involved in the Materials Science 
and Engineering Department, I would like to express my thanks for their constructive 
suggestions and cheerful assistance during the course of this work. I also 
appreciate the experimental help rendered by Adam Bogan, Joseph Puglisi, Matt 
Guyot, Robert Pekrul and Andrew Gartskiewicz. 

I wish to acknowledge the NSF Engineering Research Center for Particle 
Science and Technology at the University of Florida for providing financial support 
(through Grant # EEC-94-02989) and a stimulating interdisciplinary research 
environment. 

Last, but not the least, I would like to acknowledge my wife Anuradha for her 
invaluable assistance during the preparation of this manuscript. 



TABLE OF CONTENTS 

ACKNOWLEDGMENTS " 

ABSTRACT xii 

CHAPTERS 

1 INTRODUCTION 1 

Adsorption of Polymers at Solid/Liquid Interface 1 

Polymer Adsorption 1 

Applications 3 

Polymer Adsorption Mechanisms 3 

2 BACKGROUND 9 

Introduction 9 

Source of Isolated Hydroxyls on Dolomite 10 

Adsorption Mechanism of PEO on Dolomite Samples 10 

Deficiencies in the Proposed Adsorption Mechanism of PEO 10 

Scope of the Present Study 13 

3 EXPERIMENTAL 14 

Materials 14 

Oxide Samples 14 

Silicate Samples I 4 

Polymers 19 

Other Chemicals 19 

Methods 19 

Chemical Composition 19 

Particle Characterization 21 

PEO Characterization 23 

Surface Chemical Characterization 26 

AFM Studies 27 

Flocculation Studies 28 

Adsorption Studies 30 

4 FLOCCULATION AND ADSORPTION STUDIES ON OXIDES 33 

Introduction 33 



in 



Flocculation Studies 33 

Effect of Polymer Molecular Weight 33 

Effect of Dosage 33 

Effect of Floe Detection Technique 36 

Electrokinetic Studies 38 

Effect of pH on Flocculation of Oxides with PEO 38 

Adsorption Studies 45 

Adsorption Kinetics of PEO on Oxides 45 

Adsorption Isotherms of PEO on Oxides 45 

Adsorption Mechanism of PEO 52 

Effect of Negatively Charged Surface 53 

Effect of Hydrated Counter-Ions 53 

AFM Studies 54 

Role of Specific Surface Binding Sites in PEO 

Adsorption on Silica 62 

ROLE OF SURFACE ACIDITY OF OXIDES IN PEO ADSORPTION . . 63 

Introduction 63 

Accessibility of Surface Sites to PEO Molecules 63 

Concentration of Surface Hydroxyls 64 

Heat of Wetting of Oxides 66 

Nature of Surface Hydroxyls 66 

Point of Zero Charge of Oxides 66 

Correlation between Heat of Wetting and pzc of Oxides 68 

Role of Bronsted Acidity in PEO Adsorption 70 

Relation Between Type of Oxide and its Point Of Zero Charge . 71 
Adsorption and Flocculation Behavior of Mo0 3 and V 2 5 

with PEO 73 

Role of Lewis Acid Sites 82 

Oxide/PEO/CCI 4 system 84 

Hematite/Starch/Water System 84 

CHARACTERIZATION OF PEO BINDING SITES 

ON OXIDE SURFACES 86 

Introduction 86 

Surface Hydroxyls on Oxides 88 

Isolated Hydroxyls and PEO Adsorption 92 

Effect of Heat Pretreatment 94 

Adsorption of PEO on Heat Treated Samples 97 

Characterization of Surface Acidity of Oxides 101 

DRIFT Spectra of Adsorbed Pyridine on Oxides 101 

Acidity of Silanol Groups .104 



IV 



Surface Analysis of Silica and Adsorption of PEO 104 

Specificity of Hydrogen-Bonding of Isolated Silanols 104 

Effect of pH 107 

7 ADSORPTION AND FLOCCULATION BEHAVIOR OF SILICATES 

WITH PEO 109 

Introduction 109 

Adsorption and Flocculation Studies 109 

Flocculation of Silicates 109 

Effect of Flocculant Dosage 110 

Adsorption Studies on Silicates 113 

AFM Studies of Adsorbed Molecules on Tremolite and Augite 122 

Surface Characterization of Silicate Minerals 135 

Correlation between Isolated Hydroxyls and Adsorption 138 

Adsorption Mechanism(s) of PEO on Silicates 140 



8 CONCLUSIONS AND FUTURE WORK 146 

Summary 146 

Suggestions for Future Work 150 

REFERENCES 152 

BIOGRAPHICAL SKETCH 161 



LIST OF FIGURES 
Figure Page 

1.1. Conformation of the adsorbed polymer molecule 2 

1.2. Schematic illustrating steric stabilization of particles 4 

1.3. Schematic of bridging flocculation of particles 5 

1.4. Schematic illustrating the selective flocculation process 6 

3.1. Crystal structures of silicate samples 18 

3.2. Size distribution of PEO samples 25 

3.3. Calibration curves for analysis of PEO in solution 32 

4. 1 . Flocculation behavior of oxide samples as a function of 

molecular weight of PEO (dosage 0.5 mg/g; pH = 9.5) 34 

4.2. Flocculation behavior of oxide samples as a function of 

dosage of PEO of MW 8,000,000 at pH 9.5 35 

4.3. Bed volume of silica sediment as a function of PEO dosage 
(MW=5,000,000) at pH9.5 37 

4.4. Electrokinetic behavior of oxides as a function of pH 

(I = 0.03 kmol/m 3 ) 39 

4.5. Flocculation behavior of silica samples as a function of pH 

(PEO MW = 5,000,000; dosage = 0.5 mg/g) 41 

4.6. Flocculation behavior of silica A as a function of PEO 

molecular weight at different pH (PEO dosage = 0.5 mg/g) 43 

4.7. Electrokinetic behavior of silica A with and without PEO 

(1=0.03 kmol/m 3 ) 44 

4.8. Equilibrium adsorption time for PEO on oxides 

(PEO MW = 8,000,000; pH = 9.5) 46 

vi 



4.9. Adsorption isotherms for oxide-PEO system at pH 9.5 

(PEO MW = 5,000,000) 47 

4.10 Adsorption isotherms for oxide-PEO system at pH 3.0 

(PEO MW = 5,000,000) 48 

4.11 Adsorption isotherm of PEO (MW = 8,000,000) on silica A at pH 9.5. .51 

4.12. AFM image of the bare silica surface 55 

4.13. AFM Tapping Mode topographic image of adsorbed PEO 

(MW = 5,000,000 at pH 3.0) 56 

4.14. AFM Tapping Mode topographic image of adsorbed PEO 

at pH 9.5 after 1 hour of desorption 58 

4.15. AFM Tapping Mode topographic image of adsorbed PEO 

at pH 9.5 after 2 hour of desorption 59 

4.16. Effect of pH on interparticle forces between silica sphere 

and a flat plate with and without PEO (MW = 5,000,000) 61 

5.1. Schematic of Bronsted acid sites 69 

5.2. Heat of wetting of oxides as a function of their point of zero charge. 
(After [Hea65]) 70 

5.3. Electrokinetic behavior of Mo0 3 and V 2 5 suspensions as 

function of pH 75 

5.4. Adsorption isotherms of PEO on Mo0 3 and V 2 5 suspensions 

(PEO MW = 5,000,000; pH= 3.0) 76 

5.5. Saturation adsorption density of PEO (MW= 5,000,000) at pH 3.0 

as a function of the point of zero charge of oxides 77 

5.6. Flocculation behavior of Mo0 3 and V 2 5 as a function of 

molecular weight (dosage = 0.5 mg/g at pH 3.0) 79 

5.7. Flocculation of Mo0 3 and V 2 5 as a function of PEO dosage (pH 3.0). 80 

5.8. Schematic showing a Lewis acid site 84 

6. 1 . Schematic showing surface hydroxylation on various faces of anatase. 88 

6.2. DRIFT spectra of oxides in the hydroxyl region 90 

vii 



6.3. DRIFT spectra of silica B and hematite in the hydroxyl region 91 

6.4. DRIFT spectra of heat treated silica samples 96 

6.5. Schematic of i) amorphous silica surface showing the ring structure 

and ii) influence of surface curvature on H-bonding 97 

6.6. DRIFT spectra showing effect of heat treatment on the surface 
hydroxylation of alumina A 99 

6.7. Adsorption isotherms of PEO for heat treated oxides at pH 9.5 

(PEO MW =5,000,000) 100 

6.8. DRIFT spectra of pyridine treated Mo0 3 , V 2 5 and Si0 2 samples. ... 104 

6.9. Plot of the change in frequency of isolated silanols against the 
specific heat of adsorption for several vapors adsorbed on silica 
surface (data after [Kis65] and And [65a] 107 

7. 1 . Flocculation of silicates as a function of molecular weight of PEO 
(pH=9.5; dosage=1mg/g) 110 

7.2. Flocculation behavior of augite as a function of PEO dosage 

atpH9.5 113 

7.3. Flocculation behavior of tremolite as function of PEO dosage 

at pH = 9.5 115 

7.4. Kinetics of PEO adsorption on tremolite and augite 116 

7.5. Adsorption isotherms of PEO on chain and orthosilicates at pH 9.5 
(PEO MW = 5,000,000) 117 

7.6. Adsorption isotherms of PEO on clays at pH 9.5 

(PEO MW = 5,000,000) 118 

7.7. AFM image of bare tremolite surface 124 

7.8. AFM image of adsorbed PEO on tremolite 125 

7.9. AFM Friction image of adsorbed PEO on tremolite 126 

7.10. Histogram of parking area of PEO molecules on tremolite 127 

7.1 1 AFM image of bare augite surface 131 



VIII 



7.12. AFM image of adsorbed PEO on augite 132 

7.13. AFM Friction image of adsorbed PEO on augite 133 

7.14. Histogram of parking area of PEO molecules on augite 134 

7.15. DRIFT spectra of chain and layered silicates in the hydroxyl region. . . 137 

7.16. DRIFT spectra of orthosilicate minerals in the hydroxyl region 138 

7. 1 7. Adsorption of PEO on clays by interaction of the ether oxygen of PEO 
with the hydration shell of the exchangeable ion (Bronsted acid site). 143 



IX 



LIST OF TABLES 

Table Page 

2.1. Correlation between the intensity of the isolated hydroxyl groups on 
dolomite samples with flocculation and the saturation adsorption 
density of PEO of 5,000,000 MW [Beh93a] 11 

3.1. Oxide samples and their sources 15 

3.2. Structural Units Observed in Crystalline Silicates (After [Kin76]) 16 

3.3. Silicate samples selected and their idealized chemical composition. ..17 

3.4. Characteristics of the Polymers used in this study 20 

3.5. Physical Characteristics of the oxide samples 22 

3.6. Physical characteristics of the silicate samples 24 

3.7. Dimensions of the flocculation cell 29 

4.1. Isoelectric Point of Oxides determined by Electrokinetic Studies 40 

4.2. Saturation adsorption density of PEO (MW = 5,000,000) on oxide 
samples at different pH 49 

5.1. Concentration of surface hydroxyl groups [And82] 65 

5.2. Heat of Wetting values for oxides in water [Che59; Hea65] 67 

5.3. Probable ranges of pzc of different types of oxides [Par65] 73 

5.4. Dissolution behavior of Mo0 3 and V 2 5 powders and the 

adsorption of dissolved ions on other oxides 82 

6. 1 . Surface hydroxyls on different oxides and the saturation 

adsorption density of PEO 94 



6.2. Infrared bands of pyridine in the 1400-1700 cm" 1 

region of the spectrum 1 03 

7. 1 . Critical PEO molecular weight for flocculation of the silicate 

minerals at pH 9.5 112 

7.2. Saturation adsorption density of PEO (MW = 5,000,000) 

on chain and orthosilicates 120 

7.3. Estimated surface areas from Hg-porosimetry and the calculated 
saturation adsorption densities for tremolite and augite 122 

7.4. Type of hydroxyl groups on silicates along with PEO 

saturation adsorption density 140 



XI 



Abstract of Dissertation Presented to the Graduate School 
of the University of Florida in Partial Fulfillment of the 
Requirements for the Degree of Doctor of Philosophy 



ADSORPTION MECHANISM(S) OF POLYETHYLENE OXIDE) ON 
OXIDE AND SILICATE SURFACES 



By 

SHARAD MATHUR 

December, 1996 



Chairman: Dr. Brij M. Moudgil 

Major Department: Materials Science and Engineering 



The surface modification of solids by adsorption of polymers is critical to a 
number of industrial processes and products. In a solvent medium, polymer 
adsorption on solid substrate is exploited to disperse or aggregate the particulate 
slurries. Solid/solid separations using selective flocculation technique rely on the 
specificity of the polymer for the aggregating particles. Hydrogen bonding has been 
suggested to be the primary adsorption mechanism for non-ionic polymers and 
associated with nonselectivity. However, literature survey of PEO-oxide system 
suggested polyethylene oxide) (PEO), a non-ionic polymer, to be substrate specific. 

In this study the adsorption of PEO on various oxides and silicates and their 
flocculation behavior was systematically investigated to understand the adsorption 

xii 



mechanism(s). Surface characterization of the solid substrates was performed 
through DRIFT, AFM, adsorption and electrokinetic measurements to identify the 
adsorption mechanisms. It was shown that the adsorption of PEO is substrate 
specific indicating that hydrogen bonding is strongly dependent on the surface 
chemical nature of the substrate. It was determined that strong Bronsted acid sites 
on the surface interact with the ether oxygen, a Lewis base, of PEO to induce 
adsorption and subsequently flocculation of the particles. 

In the oxide/PEO system, highly acidic oxides such as Si0 2 , Mo0 3 , and V 2 5 
strongly adsorb PEO and exhibit flocculation. On the other hand, relatively basic 
oxides with a point of zero charge (pzc) greater than that of silica such as Ti0 2 , 
Fe 2 3 , AlpQ, and MgO, did not exhibit significant adsorption of PEO. Further, 
dissolved ions and charge characteristics were shown not to affect the adsorption 
and flocculation behavior of oxides. It was revealed for the silicate/PEO system that 
the connectivity of the silicate tetrahedra is essential in generation of strong 
Bronsted acid sites capable of interacting with PEO. 

The concept of strong Bronsted sites being essential for PEO adsorption 
provided the commonality for the binding sites identified earlier, viz., the isolated 
silanols and exchangeable ions. It also showed that rather than the isolated nature 
of the surface hydroxyls their acid strength is of prime importance in interacting with 
the ether oxygen of PEO. Further, the effect of pH on adsorption of PEO and 
similar non-ionic polymers on silica could be explained within the framework of the 
established adsorption mechanism. Additionally, the results will be beneficial to 
identification/synthesis of selectively adsorbing polymers with applications in 
processing of mineral fines, controlled drug delivery systems and ultrapurification 
of fines. 

xiii 



CHAPTER 1 
INTRODUCTION 
Adsorption of Polymers at Solid/Liquid Interface 
Polymer Adsorption 

A flexible polymer molecule such as poly(ethylene oxide) (PEO) in solution 
has a dynamically changing conformation which can be described as a random coil. 
The size of the coil is dependent upon the solvent quality, the polymer 
concentration, and characteristics of the polymer chain [Spe92]. Upon adsorption 
of the polymer molecule at the solid solution interface, the conformation of the 
polymer may change from that in the solution state. The equilibrium conformation 
is a compromise between the enthalpic factors (which tend to maximize the 
segment/surface contacts) and entropic ones (trying to maintain a thick layer with 
many degrees of freedom). The adsorbed polymer conformation is described in 
terms of trains, loops and tails (see Figure 1.1). The segments bonded to the 
surface comprise the trains, the segments between the trains constitute the loops 
and the tails are the free ends of the polymer molecule. In principle, the trains, 
loops and tails can be distinguished by the difference in their mobility by techniques 
such as Small Angle Neutron Scattering, Nuclear Magnetic Resonance and Electron 
Spin Resonance. 



3 
Applications 

Adsorbed polymer molecules at the solid/solution interface play a very 
important role in many industrial and biological products and processes. For 
instance, polymers are used as dispersants or flocculants to stabilize or aggregate 
particulate slurries, respectively. 

In order to disperse particles, a complete surface coverage of the polymer 
molecules is required in a good solvent medium so that interpenetration of the 
adsorbed polymer layers during interparticle collision leads to a net repulsive force. 
The stabilization of a suspension by this mechanism is termed as steric stabilization 
and is schematically illustrated in Figure 1.2. Dispersion of particles is important in 
ceramic and mineral processing, formulation of inks and paints, cosmetics, 
pharmaceutical and food industry. 

A partial coating of the polymer on a particle may lead to its adsorption on a 
bare surface of the other colliding particle. This is the origin of bridging flocculation 
illustrated schematically in Figure 1.3. The flocculation of particles forms the basis 
for both solid/liquid and solid/solid separations. The flocculation phenomenon, 
when it is confined to only specific type of particles in a multi-component particulate 
systems is termed selective flocculation (see Figure 1.4). 
Polymer Adsorption Mechanisms 

In order to optimize the effectiveness and efficiency of the polymer as a 
dispersant/flocculant, it is important to understand the underlying adsorption 
mechanism(s). This can be illustrated by way of example of the selective 
flocculation process. The process comprises of (1) dispersion of the fine particles 




Figure 1 .2. Schematic illustrating steric stabilization of particles. 






Figure 1 .3. Schematic of bridging flocculation of particles. 



6 



uu o o °oo 
o © J~ o 

o o o n u ® A o 

°e° ° <?8% o° 
o @ <g°o @ 

o ° ° o ®o ° 
o © Oo o 



Dispersion 



+ Polymer 

► 



O o° o^°o 
o°6oWo° 

X°°o; 



o 



Selective Flocculation 




O o O O 

o°o ° °o° 
o °o ° <&o 

°o° g °°„° o 
c> o o° o 



Separation of Floes by Sedimentation or Flotation 



Figure 1.4. Schematic illustrating the selective flocculation process. 



7 
(2) selective adsorption of the polymer on the flocculating component and formation 
of the floes, (3) floe growth which is generally achieved by conditioning at low shear, 
and (4) floe separation either through sedimentation/elutriation/sieving or flotation 
followed by cleaning of floes by repeated redispersion and flocculation, if necessary. 
Among these steps, the second step, i.e., selective adsorption of the polymer is the 
most difficult to control and by far is the limiting step to the success of the selective 
flocculation technique [Att87;Mou91]. 

Adsorption of the polymer on a particular surface is the result of the 
interactions between the functional groups and the binding sites on the surface. 
Flocculants are high molecular weight polymers which can adsorb through a number 
of mechanisms such as van der Waals, electrostatic interactions, hydrogen 
bonding, hydrophobic interactions and chemical bonding. Clearly, the surface 
property chosen (hydrophobicity, hydrogen bonding, surface charge or chemical 
bonding) must represent the one with the greatest difference between the 
flocculating and non-flocculating particles. 

Hydrogen bonding has been accepted as a ubiquitous mechanism for 
adsorption of polymers. This generalization was borne out of the fact that the oxide 
surfaces are rich in hydroxyl groups and the polymers contain a functional group 
such as the ether oxygen, or alcohol, or amide groups capable of hydrogen 
bonding. Recent investigations for the starch-hematite system, however, showed 
that the adsorption mechanism is not hydrogen bonding as was believed but a 
specific interaction between the Fe sites on the surface and the functional groups 
of the polymer [Pra91 ;Wei95]. 



8 
The aim of the present study was to understand the adsorption 
mechanism(s) of PEO on oxide and silicate surfaces since the molecule was 
reported to adsorb onto silica [Rub76;Che85;Kil84;Kok90] and certain silicates 
[Sch85;Stan90;Hog85] and not on other oxides such as hematite and alumina 
[Kok90;Sha86]. The elucidation of the adsorption mechanism(s) of PEO was 
expected to provide a unifying explanation for the experimental observations 
documented in the literature on PEO adsorption at oxide(silicate)-solution interface. 
This is also expected to lead to guidelines for identification/synthesis of selectively 
adsorbing polymers. 



CHAPTER 2 
BACKGROUND 

Introduction 
Previous studies in the silica/PEO and silica/poly(vinyl alcohol) (PVA) 
systems suggested the isolated silanols to be the principal adsorption sites for PEO 
on silica [Rub76;Tad78;Che85;Kha88]. The extension of this concept to isolated 
hydroxyls was examined for the dolomite-apatite system by Behl and Moudgil 
[Beh93d]. The DRIFT spectra of apatite and dolomite revealed that isolated 
hydroxyl groups are exclusive to the dolomite A surface, whereas hydrogen-bonded 
hydroxyl groups are present on both apatite and dolomite surface [Beh93a]. These 
investigators hydpothesized that any polymer capable of hydrogen bonding such as 
polyacrylic acid (PAA), polyacrylamide (PAM) and polyethyleneoxide (PEO) 
therefore should be capable of flocculating the two materials. Among these, PEO 
being a weak flocculant [Sch87], only the material with stronger interactions with the 
polymer molecules may be expected to flocculate. Experiments performed with 
5,000,000 MW PEO revealed that irrespective of the amount of polymer added, 
flocculation of apatite was not observed. On the other hand, instantaneous 
flocculation of dolomite occurred justifying the assumption of a specific interaction 
of the ether oxygen of PEO with the isolated hydroxyls on dolomite [Beh93d]. 



10 

The importance of the isolated hydroxyl groups was further shown through 
a correlation between their intensity on dolomite samples collected from different 
sources and the adsorption and flocculation with PEO (see Table 1) [Beh93d]. The 
dolomite samples showed similar bulk chemical composition, and it was even 
possible to separate one dolomite from another dolomite provided one of the 
dolomite could be flocculated with PEO [Mou95a]. 
Source of Isolated Hydroxyls on Dolomite 

The fact that some dolomite showed the isolated hydroxyls while the others 
did not was unexplained [Beh93d;Mou95a]. A detailed characterization of the 
dolomite samples, described below, was attempted by Moudgil et al. [Mou95b] to 
determine the cause of the isolated OH on some dolomite samples and its absence 
on others. DRIFT and X-ray Diifraction (XRD) studies revealed the presence of a 
coating of palygorskite clay on the flocculating dolomite samples. 
Adsorption Mechanism of PEO on Dolomite Samples 

Hoghooghi [Hog85] has shown that PEO is an excellent flocculant for 
palygorskite. The mechanism of adsorption of PEO on dolomite A is expected to 
be similar to the hydrogen bonding mechanism proposed for silica. The isolated OH 
on dolomite A may act as proton donor in hydrogen bonding to ether oxygen of 
PEO. 

Deficiencies in the Proposed Adsorption Mechanism of PEO 

The correlation between saturation adsorption density of PEO and 
flocculation of dolomite samples with the presence of isolated hydroxyls suggested 
that the assumption of isolated hydroxyls being the principal adsorption sites for the 



11 



Table 2.1. Correlation between the intensity of the isolated hydroxyl groups on 
dolomite samples with flocculation and the saturation adsorption density of PEO of 
5,000,000 MW [Beh93a]. 



Dolomite 


Flocculation 
(dosage 1 mg/g) 


Saturation 

adsorption 

density, mg/m 2 


Intensity of 

Isolated OH at 

3619 cm 1 


A 


98.5 


2.18 


High 


B 


92.5 


1.93 


High 


C 


69.3 


1.16 


Medium 


D 


72.5 


1.19 


Medium 


E 





0.88 


None 


F 





0.39 


None 



12 
ether oxygen of PEO is correct. However, the presence of isolated OH on dolomite 
samples was determined to be due to the coating of palygorskite clay. The 
flocculation behavior of clays such as montmorillonite and palygorskite has been 
extensively studied by Scheiner and co-workers [Sch86;Sch87;Bro89]. These 
investigators attributed the adsorption mechanism of PEO to a hydrogen bonding 
mechanism involving the water shell around the exchangeable cations on the clays 
[Sch86;Bro89], Although the isolated hydroxyls constitute the surface of other 
oxides such as alumina and hematite yet flocculation of these oxides was not 
observed [Kok90]. The presence of isolated hydroxyls on the surface of these 
oxides has been shown through vibrational spectroscopy [Hai67;Tsy72;Mor76]. 
Additionally, other possible mechanisms for adsorption of PEO suggested in the 
literature e.g. electrostatic interactions with a positively charged surface, and 
complex binding with adsorbed ions such as K + , Cd + , Mg + etc. [Bai76; Kje81; Ana87; 
Beh93d;Pra95] have not been examined in detail to explain the flocculation 
behavior of various oxides with PEO. 

Thermodynamically, the overall free energy of the polymer adsorption 
process must be negative. In addition to the enthalpic factors such as the 
segment/surface and water-surface interactions the overall entropy changes 
associated with the adsorption process are also important. In fact, the lack of PEO 
adsorption on alumina and hematite was suggested to be due to the lack of 
accessibility of PEO molecules to the surface sites [Kok90]. The entropy factor was 
earlier invoked by Greenland to explain the unreactivity of aluminol and silanol 
groups with PVA [Gre72a,b]. Thus the adsorption mechanism(s) need to consider 



13 
the entropy contribution to polymer adsorption. Further, the adsorption 
mechanism(s) must also be consistent with the observed decrease in adsorption of 
non-ionic polymers, such as PEO and PVA, with pH [Rub76;Che85;Tad78;Kha88]. 

In order to further understand the adsorption mechanism of PEO and similar 
non-ionic polymers, a systematic study involving surface-chemical characterization 
of various oxides and different types of silicates was undertaken. The reason for 
selecting a variety of substrates was to examine all the possible mechansims 
reported in the literature including the accessibility factor, and establish the 
predominant mecahnism of PEO adsorption on oxides and silicates. 

Scope of the Present Study 

The specific objectives of the present investigation are as follows: 

1 . Establish the mechanism(s) of PEO adsorption on oxides. Specifically, it is 
proposed to understand the role of different type of surface sites in 
interaction with the ether oxygen of PEO. 

2. Examine the effect of pH on adsorption of non-ionic polymers such as PEO. 

3. Evaluate the role of surface accessibility in polymer adsorption. 

4. Examine the adsorption mechanism(s) of PEO on mixed oxides such as 
silicates based on better understanding of the same on simple oxides. 



CHAPTER 3 
EXPERIMENTAL 
Materials 
Oxide Samples 

The oxide samples along with their sources are listed in Table 3.1. These 
were selected so as to encompass a range of acidic to basic surfaces. The oxide 
samples were used as received except V 2 5 and Silica B which were wet ground 
to obtain -400 mesh (<38 urn) fraction. 
Silicate Samples 

The silicate samples chosen were representative of the different classes of 
silicates. The characteristic structural features of these classes are summarized 
in Table 3.2. The minerals selected and their idealized chemical compositions are 
listed in Table 3.3. Although the Si/O ratio in palygorskite is 2.75, which is 
characteristic of the amphiboles, the arrangement of the chains is such that the clay 
is considered to be a pseudo-layered silicate [Gri68]. The crystal structures of the 
silicate samples are shown in Figure 3.1. The as received samples from Ward's 
Natural Establishment Inc., NY, were crushed in a Chipmunk crusher and then 
pulverized and subsequently sieved to yield different size fractions. The -400 mesh 
fraction of the silicate samples was used in this study. 



14 



Table 3. 1 . Oxide samples and their source. 



15 



Oxide Sample 


Code 


Source 


Silica 


(A) 


Geltech Inc., FL 


(B) 


IMC-Agrico, FL 


Titania 




Alfa 


Hematite 




Alfa 


Alumina 


(A) 


Sumitomo 


(B) 


Alcoa 


Magnesia 




Mallingcrockdt 


Molybdenum 
Oxide 




Alfa 


Vanadium 
Oxide 




Alfa 



16 



Table 3.2. Structural Units Observed in Crystalline Silicates (After [Kin76]). 



Oxygen-Silicon 
Ratio 


Silicon-Oxygen 
Groups 


Structural Units 


Type 


2 


Si0 2 


Three dimensional 
Network 


Framework 


2.5 


Si 4 O 10 


Sheets 


Layered 


2.75 


siAi 


Chains 


Amphiboles 


3.0 


Si0 3 


Chains 


Pyroxenes 


4.0 


Si0 4 


Isolated 

orthosilicate 

tetrahedra 


Orthosilicates 



17 



Table 3.3. Silicate samples selected and their idealized chemical composition. 



Material 


Type of silicate 


Chemical Formula 


Palygorskite 


Layer 


(0H ? ) 4 (0H) ? Mg fi Si„0„. 4H,0 


Kaolinite 


Layer 


AI 4 SL0 in (0H)„ 


Tremolite 


amphibole 


Ca ? Mg,[Si„0„](OH), 


Augite 


pyroxene 


(Ca,Na)(Mg,Fe,AI)(Si,AI) ? O fi 


Almandite 


orthosilicate 


i 6^A\lpOl^v*/^p 


Topaz 


N 


AI,(Si0 4 )(0H) 


Olivine 


II 


(Mg,Fe) ? Si0 4 




(i) 





18 






(iii) 




(iv) 



Figure 3.1. Crystal structures of silicate samples: (i) isolated Si0 4 tetrahedra as 
in almandine, olivine and topaz, (ii) Si0 4 chain in pyroxene, augite, (iii) double 
Si0 4 chains in amphibole, tremolite and (iv) Si0 4 layer in kaolinite. 



19 

Polymers 

Poly(ethylene oxide) (PEO) of different molecular weights along with the 
source and calculated radius of gyration is listed in Table 3.4. These samples were 
used as received. 
Other Chemicals 

All experiments were conducted in distilled water (Dl) of specific conductivity 
less than 1 umho/cm. Potassium hydroxide (KOH) and nitric acid (HN0 3 ) used as 
pH modifiers were obtained from Fisher Scientific Co. Pyridine, used as a probe 
molecule for characterization of the surface acidity, was also procured from Fisher 
Scientific Co. 

Methods 
Chemical Composition 

The as received oxide samples were specified to be of more than 99.9% 
purity. Silica A was prepared by the sol-gel technique (Stober silica) while silica B 
was electrostatically separated from the fluorapatite and acid washed to remove the 
minor phosphate impurity. The P 2 5 content of the silica B sample was below the 
detection limit (0.05 wt.%) of the Perkin Elmer II inductively coupled plasma (ICP) 
spectrometer. Alumina B was determined by ICP to contain, on weight basis, 
99.9% Al 2 3 , 0.01% Si0 2 , 0.02% Fe 2 3 and 0.07% Na 2 0. 

The silicates samples were characterized for the purity primarily by the 
hydroxyl band of their DRIFT spectra. No quantitative analysis of the samples was 
attempted. 



20 



Table 3.4. Characteristics of the Polymers used in this study. 



Source 


Molecular Weight 


Radius of Gyration, R g 
(nm) 


Polysciences, Inc. 


8,000,000 


177 


Polysciences, Inc. 


5,000,000 


144 


Polysciences, Inc. 


4,000,000 


126 


Aldrich 


900,000 


52 


Polysciences, Inc. 


600,000 


41 


Polysciences, Inc. 


1,00,000 


14 


Polysciences, Inc. 


18,500 


6 



21 
Particle Characterization 

All the oxide and silicate powders were characterized for particle size 
distribution and surface area. 
Oxides 

Particle size distribution. The particle size distribution of the oxide samples 
was determined using Micromeritics X-ray Sedigraph 5100 and the characteristic 
diameters are presented in Table 3.5. It is observed that silica A is monodisperse. 
The silica A particles are spherical in shape, which is a characteristic of the Stober 
process to synthesize silica. Silica B, Mo0 3 and MgO particles are coarser than the 
other oxide samples. 

Surface area. The surface area was essential to compare the adsorption 
density of PEO on different substrate. The size of the high molecular weight PEO 
is such that the BET method is insensitive to the corresponding pore size, e.g., 288 
nm for 5,000,000 MW PEO molecule. The specific surface area of the oxide 
samples was, therefore, determined by mercury porosimetry (Micromeritics 
Autopore III 9420) and the results are presented in Table 3.5. 

In mercury porosimetry the raw data generated for a powder sample consists 
of the intrusion volume of mercury versus the applied pressure. The interparticle 
pores are filled first and the slope of the graph changes at the point when the 
intraparticle pores on the surface are intruded by the mercury. The pore radius and 
corresponding surface area are calculated from the assumption of a cylindrical 
geometry of the pore. The calculated surface area for non-porous spherical 
particles of silica A (1pm) is 2.85 m 2 /g which is in good agreement with the 



22 



Table 3.5. Physical Characteristics of the oxide samples 



Sample 


Average Particle Size, pm 


Specific Surface 

Area, 

m 2 /g 




die 


d 50 


d 8 4 


Si0 2 (A) 


1+0.1 


3.18 


Si0 2 (B) 


2.4 


8.5 


18.0 


1.91 


Ti0 2 


0.1 


0.3 


0.5 


11.10 


Fe 2 3 


0.3 


0.6 


1.5 


8.23 


Al 2 3 (A) 


0.2 


0.5 


0.8 


7.11 


Al 2 3 (B) 










MgO 


0.8 


1.7 


6.8 


3.38 


Mo0 3 


3.1 


6.4 


8.2 


1.33 


v,o< 


0.2 


0.6 


1.0 


7.13 



23 

experimental value of 3.18 m 2 /g. Similarly, the size distribution of the other oxide 
samples and the surface area values indicate that the particles are essentially non- 
porous. 
Silicates 

Particle size distribution. The particle size distribution of the ground -400 
mesh fraction (< 38 urn) of tremolite and augite determined by Micromeritics X-ray 
Sedigraph 5100 is shown in Table 3.6. It is revealed that the silicate particles are 
coarser than the oxide samples. 

Surface area. The specific surface area of the silicate minerals was 
determined by Autosorb Micromeritics ASAP 2000 system. The surface area was 
estimated by the BET method using adsorption and desorption of nitrogen and is 
presented in Table 3.6. The relatively higher specific surface areas for tremolite 
and augite indicate that the samples have a significant amount of porosity 
associated with them. The Hg-porosimetry, as in the case of the oxides, was, 
therefore, used to determine the effective surface area for PEO adsorption for 
tremolite and augite. 
PEO Characterization 

The PEO samples received were granular in appearance and the molecular 
weight specified by the manufacturer is an average value. In order to determine the 
polydispersity of the high molecular weight fractions light scattering on PEO 
solutions was performed using Brookhaven Bl 90. The size measured is the 
hydrodynamic diameter (2xR g z ) of the polymer molecules in solution. The results 
presented in Figure 3.2 reveal that the polymers are polydisperse in nature. 



24 



Table 3.6. Physical characteristics of the silicate samples. 



Sample 


Average Particle Size, pm 


Specific Surface Area, 
m 2 /g 


d 16 


d 50 


d 84 


Hg-P* 


BET 


Tremolite 


6.0 


11.0 


32.0 


8.81 


13.88 


Augite 


5.0 


15.0 


30.0 


3.51 


4.11 


Kaolinite 


<38.0 


- 


15.52 


Palygorskite 


- 


117.0 


Almandite 


- 


0.79 


Olivine 


- 


1.33 


Topaz 








- 


4.54 



Hg-P' = mercury porosimetry 






25 



100 



CD 



CD 
> 

=3 

E 
o 




150 200 

Hydrodynamic Diameter (nm) 



250 



300 



Figure 3.2. Size distribution of PEO samples. The molecular weights are indicated 
in the legend. 



26 
Surface Chemical Characterization 

The determination of the adsorption mechanism of PEO involved a 
knowledge of the surface chemical groups and charge characteristics of the 
samples. 
Infrared studies 

Samples used for FT-IR analysis were prepared by mixing a 0.05g of 
vacuum dried sample with about 0.75 g of potassium bromide (KBr). The mixture 
was then filled in a sample cup mounted on a diffuse reflectance stage. The IR 
spectra was taken using a Nicolet 740 spectrometer. The beam was aligned for 
every sample to yield a maximum signal and the baseline was adjusted using the 
software. The evolution of the DRIFT spectra as a function of temperature was 
performed on Nicolet 60SX spectrometer using the heating stage. 

The acidity of the surface chemical groups on the oxides was probed by 
pyridine adsorption which involved soaking 1g of the sample in 25 ml liquid pyridine 
for 1 hour. The samples were air-dried under a fumehood and subsequently kept 
in a vaccum oven at 60°C for 12 hours. The samples were then mixed with KBr 
prior to obtaining the DRIFT spectra. 
Electrokinetic studies 

Zeta potential of the oxide samples was measured using the Laser Zee 
Meter (Pen Kern Model 501) to determine their isoelectric point. 50 mg of the 
powder sample was suspended in 100 ml of .03 kmol/m 3 KN0 3 solution except for 
Mo0 3 and V 2 5 which exhibited ionic strength in excess of .03 kmol/m 3 . 



27 
For these two oxides, the centrifuged suspension from a 2 wt.% slurry was used as 
such to determine the zeta potential of the suspended solids. The Laser Zee Meter 
was also used to determine the zeta potential of PEO coated silica samples. 
AFM Studies 

The Atomic Force Microscope (AFM) studies were conducted to characterize 
the conformation of the adsorbed polymer layer at the solid/solution interface. The 
samples were imaged in a liquid cell where the suspension conditions such as the 
pH and ionic strength can be simulated. The samples used for the AFM study were 
as-received fused silica plates from Herasil Amereus and polished samples of 
augite and tremolite rocks which were received from Ward's Natural Establishment 
Inc., NY. The final polishing of the silicate samples was done using 0.03 urn iron 
oxide suspension from Buehler. 

The AFM was used in both the contact and tapping modes to image the 
microstructure at the solid solution interface. The interpretation of the images was 
facilitated by the image analyzer software of the AFM. 

The contact mode was also used to obtain force/distance profiles between 
a glass sphere of 10-40 urn diameter attached to the cantilever and the fused silica 
plate with and without the adsorbed polymer. These profiles were obtained by 
suspending the x-y raster motion of the piezoelctric and measuring the deflection 
of the cantilever as it approached the surface. The cantilever used had a force 
constant of 0.30 N/m. Thus the measured deflection values were converted to the 
corresponding force values as a function of the separation distance. 



28 
Flocculation Studies 

These were conducted to macroscopically evaluate the effect of PEO 
adsorption on the particles. 
Flocculation apparatus 

The flocculation of particulate suspensions has been shown to be sensitive 
to the type of agitation used [Hog85]. Hence it is necessary to maintain uniform, 
hydrodynamic conditions in all the flocculation experiments. The mixing unit 
employed in this study is based on the standard tank design [Dir81], 
The dimensions of the mixing tank are listed in Table 3.7. A 150 ml beaker fitted 
with removable plexiglas baffles of appropriate dimensions was used for flocculation 
tests. A stainless steel turbine impeller with four blades mounted on a variable 
speed motor was employed to agitate the sample. 
Flocculation procedure 

Material suspension of pulp density 2g/100 ml was prepared in Dl water and 
aged for one hour. The maximum pH variation after aging was determined to be 0.2 
pH units. After aging, the suspension was agitated at 1 100 rpm, for 240 seconds 
while the pH was adjusted to the desired value. The suspension was sonicated for 
30 seconds at a setting of 50 W to ensure complete dispersion, and a pre- 
determined amount of PEO was added. The agitation was continued at 1 100 rpm 
for 120 seconds within which the flocculation was observed to be complete. The 
formation of floes was evaluated in a sedimentation column while the quantification 
of floe formation was obtained via floe separation over a 400 mesh screen. 



29 



Table 3.7. Dimensions of the flocculation cell. 



Flocculation Cell Part 


Dimension, cm 


Tank diameter 


6.2 


Impeller height from tank bottom 


0.5 


Impeller blade width 


1.6 


Liquid height 


4.6 


Baffle width 


4.0 






30 
Polymer Solution Preparation 

Polymer solutions were prepared by mixing 0.25 g granular polymer with 500 
ml Dl water to yield a 500 ppm solution. The solution was stirred for 16h at 500 rpm 
and covered to avoid exposure to ultraviolet radiation which decomposes the 
polymer. The polymer solution was prepared fresh every day for the experiments 
since changes were observed in PEO adsorption when stored for more than one 
day [Mou92]. 
Adsorption Studies 

These were conducted to measure the affinity of the surface for the PEO 
molecules. The adsorption isotherms, in conjunction with the knowledge of the 
surface chemical groups, provided insight into the adsorption mechanism of PEO 
on oxide and silicate surface. 

Adsorption was carried out by contacting the polymer solution with 2 wt.% 
solids in 100 ml solution in 150 ml beakers (same as the flocculation tank). Sixteen 
beakers were simultaneously stirred on a 16-pad magnetic stirrer at 500 rpm. The 
equilibrium time for adsorption was first determined by studying the kinetics of 
polymer adsorption with a high polymer dosage (10mg/g solids or 200 ppm). After 
equilibration the sample was centrifuged at 15,000 rpm for 10 minutes and the 
supernatant withdrawn. The residual PEO in the solution was determined by Total 
Organic Carbon (TOC) analyzer. Adsorption was determined by the solution 
depletion method. The saturation adsorption density, which is the maximum 
possible amount of polymer on the surface, was obtained by fitting Langmuir 
equation for high concentrations [Att91, Beh93]. 



31 

PEO Analysis 

The amount of residual polymer in the supernatant was determined using 
the TOC (Shimadzu). Calibration curves were obtained for PEO of MW 5,000,000 
and 8,000,000 by using standard solutions of PEO (see Figure 3.3) and correcting 
the measured values for the initial carbon content in the supernatant. The inorganic 
carbon was minimized by acidifying the PEO solutions with 85 wt.% phosphoric acid 
and sparging the solution with carbon-free air from gas cylinder. The concentration 
value measured had a cumulative variance of less than 1%. 



32 



35000 



30000 - 



25000 



£ 20000 
o 

Q) 

-I— " 

a> 15000 

T3 

C 
13 

03 

| 10000 



5000 







I I I 


I i 

PI 
/ 

/ 

/ 






o MW = 5,000,000 
□ MW = 8,000,000 




I I I 


i i 









50 100 150 

PEO concentration, ppm 



200 



250 



Figure 3.3. Calibration curves for analysis of PEO in solution. 



CHAPTER 4 
FLOCCULATION AND ADSORPTION STUDIES ON OXIDES 

Introduction 

In this Chapter the flocculation behavior of common oxides such as silica, 
titania, hematite, alumina and magnesia with PEO and the related adsorption 
studies are discussed to identify the underlying adsorption mechanism(s). 

Flocculation Studies 
Effect of Polymer Molecular Weight 

The adsorption of a relatively high molecular weight polymer on the substrate 
generally implies the possibility of floe formation. However, there exists a critical 
molecular weight beyond which one can detect flocculation [Beh93a]. At lower 
molecular weight the polymer may act as a dispersant. 

The flocculation behavior of all the oxides as a function of the molecular 
weight of PEO at 0.5 mg/g dosage is shown in Figure 4.1. It is observed that- 
irrespective of the molecular weight of PEO no flocculation of any oxide except silica 
A is observed. The critical molecular weight of PEO at which flocculation of silica A 
occurred was found to be 8,000,000. 
Effect of Dosage 

It is observed from Figure 4.2 that silica flocculation exhibits a maximum as 
a function of PEO dosage, whereas the other oxides did not flocculate in the dosage 



33 



34 



100 



T3 

CD 

=3 

o 

O 

o 



c 

3 

o 
E 
< 



80 



60 



40 



20 







silica A 



(BOO 



-0- 




Silica B, Titania, Hematite, Alumina, Magnesia 



0e+0 2e+6 4e+6 6e+6 

Polymer Molecular Weight 



8e+6 



1e+7 



Figure 4.1 . Flocculation behavior of oxide samples as a function of 
the molecular weight of PEO (dosage = 0.5 mg/g; pH = 9.5). 









35 



-a 

O 

o 
_o 

LL 

C 
=3 
O 

E 
< 



100 



80 



60 



40 



20 - 








Silica B, Titania, Hematite, Alumina, Magnesia 






2 4 6 8 

Flocculant Dosage, mg/g 



10 



12 



Figure 4.2. Flocculation behavior of oxide samples as a 
function of dosage of PEO of MW 8,000,000 at pH 9.5. 



36 
range examined. The existence of a maximum in flocculation followed by 
restabilization is in agreement with the bridging mechanism of flocculation. In silica- 
PEO system a similar flocculation behavior with dosage has been reported in the 
past by Rubio and Kitchener [Rub76] and Cheng [Che 85] with PEO of 5,000,000 
MW. 
Effect of Floe Detection Technique 

It must be noted that silica particles are 1 urn in size and floes are separated 
using a 400 mesh (37pm) screen. It is, therefore, likely that smaller floes formed 
with a lower molecular weight flocculant may not be detected by sieving. Thus, 
determination of the critical molecular weight is strongly influenced by the floe 
detection technique. 

In order to observe flocculation with 5,000,000 MW PEO settling tests were 
conducted on silica A and the results are shown in Figure 4.3. It is seen that the 
trend of bed volume with dosage is similar to that shown in Figure 4.2 indicating floe 
formation in the system. Koksal et al [Kok90] and Shah [Sha86] reported no 
measurable flocculation of hematite and alumina with PEO using settling tests to 
detect the onset of flocculation. 

The lack of flocculation of silica B at pH 9.5 is not unexpected. Koksal et al 
[Kok90] observed the flocculation of quartzite only near the isoelectric point of 2.5 
while no flocculation was detected beyond pH 4.0. Rubio and Kitchener [Rub76] 
observed that the precipitated silica was virtually non-flocculable at higher than pH 
8.0 while the heat treated silica was only slightly flocculated when the pH was raised 
to 9.5 from 2.0. Cheng [Che85] also observed an enhanced flocculation of silica 



37 



E 

o 

> 

"O 
<D 
CD 








60s 
120 s 
600 s 
1200 s 
3600 s 
7200 s 
10800 s 
14400 s 






12 3 4 

Polymer Dosage (mg/g) 

Figure 4.3. Bed volume of silica sediment as a function of PEO dosage 
(MW = 5,000,000) at pH = 9.5. 



38 
at pH 3.7. Thus flocculation of silica B is expected near its isoelectric point and this 
may be true for the other oxides. 
Electrokinetic Studies 

The electrokinetic behavior of the oxide samples is summarized in Figure 4.4 
The measured isoelectric point (iep) values summarized in Table 4.1 are in 
agreement with the literature values [Ree92]. 
Effect of pH on Flocculation of Oxides with PEO 

The effect of pH on flocculation of the silica samples with PEO of 5000,000 
MW is shown in Figure 4.5. In accordance with previous work flocculation of both 
the silica samples in the acidic pH range was observed. Further, a sharp decrease 
in flocculation beyond pH 3.0 was noticed indicating a decrease in the adsorption 
of PEO on negatively charged surfaces. In fact, when the pH of the flocced slurry 
of silica B was raised to 9.5 the floes begin to disappear and a dispersed 
suspension of silica particles was obtained. 

Flocculation of oxides other than silica was not observed in the pH range 2.5- 
10.0. The non-flocculation of hematite and alumina in this pH range was also 
reported by Koksal et al [Kok90]. This implies that neither neutral sites such as 
MOH (which are a maximum at the isoelectric point) nor positively charged sites 
such as MOH 2 + (which predominate below the isoelectric point) lead to adsorption 
of PEO on these oxides. Further, silica A flocculated when it exhibited a high 
negative charge, and both silica samples flocculated near the isoelectric point. Thus 
electrokinetic characteristics of an oxide sample do not necessarily govern the 
adsorption of PEO. 



39 



80 



60 



40 



> 

§ 20 

c 
ID 

o 
a. 

S 

CD 
N 







-20 



-40 



o 


silica B 


a 


silica A 


A 


titania 


V 


hematite 


O 


alumina A 





alumina B 




-60 







6 
PH 



8 



10 



12 



Figure 4.4. Electrokinetic behavior of oxides as a function of pH 
(KN0 3 = 0.03 kmol/m 3 ). 






40 



Table 4.1 . Isoelectric Point of Oxides determined by Electrokinetic Studies 



Oxide 


Isoelectric Point 


Silica A 


2.0 


Silica B 


2.0 


Titania 


3.8 


Hematite 


8.4 


Alumina A 


8.8 


Alumina B 


7.4 


Magnesia 





41 



CD 

iS 

3 
o 
o 
_o 

LL 

■—• 
C 
=S 
O 

E 
< 



100 



80 



60 



40 



20 












I 

silica A 
silica B 



-DO — DO DO 



8 



10 



PH 



Figure 4.5. Flocculation behavior of silica samples as a function of 
pH (PEO MW = 5,000,000; dosage = 0.5 mg/g). 









42 
The critical molecular weight of PEO for flocculation of silica A shifted to 
900,000 from 8,000,000 when the pH was decreased from 9.5 to 3.0 (see Figure 
4.6). Assuming the adsorption behavior of PEO to be similar at both pH values the 
increase in flocculation of silica particles with decrease in pH may be attributed to 
the reduction in the zeta potential of the silica surface allowing a closer approach 
between the particles. In such a case it is only with PEO of 8,000,000 MW that the 
zeta potential is expected to be significantly low at pH 9.5 due to the adsorbed layer 
exceeding a critical thickness of the order of the thickness of the order of the 
thickness of the electrical double layer (110 nm at l=10" 5 kmol/m 3 ). 

The electrokinetic data for silica A with and without adsorbed PEO of 18,500 
MW is plotted in Figure 4.7. It is indicated that the adsorbed layer thickness of the 
18,500 MW PEO is sufficient to result in a zero zeta potential. Thus a decrease in 
the zeta potential is not the reason for a decrease in the critical molecular weight of 
PEO for flocculation of silica suspensions with increase in pH. Rubio and Kitchener 
[Rub76] and Cheng [Che85] showed that the saturation adsorption density of PEO 
on silica decreases with increase in pH. A direct correlation between flocculation 
and saturation adsorption density was suggested by Behl and Moudgil [Beh93] for 
adsorption of PEO on several dolomite samples. Adsorption studies were therefore 
undertaken to explain the flocculation behavior of the different oxide samples and 
identify the PEO adsorption mechanism(s). 



43 



T3 

o 
is 

o 

o 
o 



c 

o 
E 
< 



100 - 




10000 



100000 1000000 

Polymer Molecular Weight 



1 0000000 



Figure 4.6. Flocculation behavior of silica A as a function of PEO 
molecular weight at different pH (PEO dosage = 0.5 mg/g). 



44 



> 

« 
c 

CD 
O 

a. 

S 

N 







-10 



-20 



-30 



-40 



-50 



-60 




O 
□ 



PEOMW 18,500 
without PEO 



PH 



10 



Figure 4.7. Eelctrokinetic behavior of silica A with and without PEO 
(1=0.03 kmol/m 3 ). 



45 
Adsorption Studies 
Adsorption Kinetics of PEO on Oxides 

The adsorption kinetics of PEO of 8,000,000 MW on the different oxide 
samples were examined and the results are presented in Figure 4.8. It is observed 
that equilibration is achieved in about 4 hours for silica A while the remaining 
samples did not exhibit any adsorption of PEO even after 24 hours. Thus further 
adsorption tests were performed with an equilibration time of 4 hours. 
Adsorption Isotherms of PEO on Oxides 

The adsorption isotherms for all the oxide samples at pH 9.5 and 3.0 with 
PEO of 5,000,000 MW are plotted in Figures 4.9 and 4.10 respectively. The 
saturation adsorption density of PEO on various oxides under the two pH levels is 
summarized in Table 4.2. Considering that an equivalent monolayer of PEO 
corresponds to a saturation adsorption density of 0.4 mg/ m 2 it is clear that oxides 
other than silica did not exhibit significant surface coverage by PEO. 

It is observed from Figures 4.9 and 4.10 that in contrast to silica A PEO 
adsorption on silica B was significantly affected by pH. The notable effect of pH on 
adsorption of PEO on silica is indicative of the flocculation behavior of the silica 
suspensions described in Figures 4.2 and 4.5. Although the saturation adsorption 
density of PEO on silica A is significant with 5,000,000 MW it is 25% lesser at the 
higher pH indicating a lesser number of the active sites and hence a lower 
probability for flocculation at pH 9.5 according to the equivalent site concept 
proposed by Behl and Moudgil [Beh93c]. Thus larger floes were obtained only in 
the acidic pH range and not at pH 9.5 (see Figure 4.5). However, with 8,000,000 



46 



2.0 



1.5 - 



E 



in 

& 1.0 

c 
o 



o 

(0 

T3 
< 



0.5 - 



0.0 



I I I 


I I 






O silica A 




^/"vOfV— <^_^ /^v _—■ /^ O— — _ 






i i i i i 


- 







10 



15 



20 



25 



30 



Time (h) 
Figure 4.8. Equilibrium adsorption time for PEO on oxides (PEO MW = 
8,000,000; pH 9.5). 









47 



2.0 



CM 



CD 

E 

-6 
9 

o 

w 

< 

CD 
E 

o 

Q_ 



c 
3 

o 
E 
< 



1.6 



1.2 



0.8 



0.4 



0.0 



o 


silica B 


□ 


alumina A 


A 


alumina B 


V 


titania 


o 


hematite 





magnesia 




-o- 



-o 



J_ 



40 60 80 100 120 
Residual Concentration, ppm 



140 160 180 



Figure 4.9. Adsorption isotherms for oxide-PEO system at pH 9.5 
(PEO MW = 5,000,000). 



48 



2.0 



1.6 - 



CM 



f 

E 
■o 

CD 

* 1-2 
o 

co 
< 

E 

I 0.8 



c 

3 
O 

E 
< 



0.4 



0.0 



I I 


I I ! 




I I 


i 




o 


silicaA 








a 


silica B 










A 


alumina A 










V 

O 




alumina B 

titania 

hematite 






- 




o 


magnesia 




— 








- °o ° 






<d 


o 


oB D 


□ 




o 




i 








- 


S_jQO-^7- 


-Q^O-iO^aii 


— Oft 1 L 





20 40 60 80 



100 120 
Residual Concentration, ppm 



140 160 



Figure 4.10. Adsorption isotherms for oxide-PEO system at pH 3.0 
(PEO MW = 5,000,000). 






49 



Table 4.2. Saturation adsorption density of PEO (MW = 5,000,000) on oxide 
samples at different pH. 



Oxide 


Saturation Adsorption Density, 
mg/m 2 


pH3.0 


pH9.5 


Silica A 


0.80 


0.65 


Silica B 


0.63 


0.17 


Titania 


<0.1 


Hematite 


Alumina A 


Alumina B 


Magnesia 







50 
MW PEO large floes were observed at pH 9.5 indicating the adsorption to be similar 
to that obtained at pH 3.0 with 5,000,000 MW PEO. The saturation adsorption 
density of 8,000,000 MW PEO at pH 9.5 from the isotherm in Figure 4.11 was 
determined to be 0.8 mg/m 2 which is similar to that obtained for 5,000,000 MW PEO 
at pH 3.0 (see Table 4.2). 

The saturation adsorption density of 0.8 mg/m 2 of PEO on silica A is twice 
that of the equivalent monolayer adsorption density. Fleer et al [Fle83] and 
Blaakmeer [Bla90] have suggested that the saturation adsorption is about 2 to 5 
times the equivalent monolayer, primarily due to compaction of the polymer 
molecule at high adsorption densities. The maximum in flocculation of silica A at 
pH 9.5 is observed at about 0.16 mg/m 2 adsorption density ( 0.4 mg/g dosage) with 
both 5,000,000 and 8,000,000 MW PEO (see Figures 4.2 and 4.3). The occurrence 
of maximum in floe formation at less than half the surface coverage (0.2 mg/m 2 ) 
may be attributed to the polydispersity of the polymer [Beh93b], 

Adsorption studies for PEO/oxide system besides silica have been reported 
only for alumina. The negligible adsorption of 5,000,000 MW PEO on alumina was 
earlier shown by Shah [Sha86]. In a recent study of adsorption behavior of 8000 
PEG on oxides and silicates it was shown that the adsorption of PEO on alumina 
was negligible [Wal96]. The presence of impurity ions such as sodium on alumina 
B also did not influence the adsorption results when it is well known that sodium 
ions complex with PEO in solution [Ana87;Bai76;Pra95] indicating that complexation 
with adsorbed surface ions is not a mechanism for PEO adsorption. 



51 



CM 



en 
E 

"d 

CD 

25 

i— 

o 
in 

< 

E 

o 
a. 



c 

3 
O 

E 
< 




20 40 60 80 

Residual Concentration of Polymer, ppm 

Figure 4.1 1 . Adsorption isotherm of PEO (MW = 8,000,000) on 
silica A at pH 9.5. 



100 



52 
Adsorption Mechanism of PEO 
It has been shown so far that positively charged sites (MOH 2 + ) or adsorbed 
ions which complex with ether oxygen in solution, and electrostatic considerations 
do not play a major role in the PEO adsorption process. Rubio and Kitchener 
[Rub76] suggested the following possible adsorption mechanisms of PEO on silica 
to explain the effect of pH. 

1. Increased repulsion of PEO from an increasingly negatively charged 
interface as the pH is increased. This is because the ether oxygen with the 
two lone pair of electrons is considered to be slightly negatively charged. 
Thus the effect of surface ionization is through a general double-layer 
phenomenon rather than the loss of binding sites. The evidence in support 
of this explanation was the slight increase in PEO adsorption at a given pH 
with electrolyte addition. 

2. The binding sites for the ether oxygen of PEO are the surface hydroxyls on 
the solid surface. It has been shown by infra-red studies that hydroxyls of 
different acid strength are present on the silica surface. The chemical 
characteristics of the surface hydroxyls have been shown to vary and it is 
possible that the most readily dissociable silanol group, i.e., the most acidic 
group are the most important one for adsorption of PEO. These sites are 
thus ionized first as the pH is increased leading to decreased adsorption of 
PEO. 

3. Hydrated counter-ions prevent PEO from approaching the surface as the 
negative charge increases with the pH. This hypothesis was also suggested 



53 
by Her [Ile75] who showed that incorporation of aluminosilicate anions into the silica 
surface enabling it to retain a negative charge in the acidic solution led to a 
decrease in flocculation. 
Effect of Negatively Charged Surface 

The non-adsorption of PEO on oxides other than silica at their isoelectric 
points indicates that explanation (1) above probably plays a minor role in adsorption 
of PEO. The repulsion of PEO molecule from a negatively charged silica surface 
resulting in decreased adsorption upon increase in pH implies that a positively 
charged surface should attract PEO which was found not to be the case as oxides 
below their iep did not exhibit significant adsorption of PEO. Also, electrokinetic 
studies showed that silica A possesses a similar charge as silica B at any pH yet 
their adsorption behavior is very different. 

The evidence provided by Rubio and Kitchener [Rub76] in favor of the 
electrostatic interactions is the slight increase in PEO adsorption with electrolyte 
addition. There is about 10% increase in saturation adsorption for precipitated silica 
in the presence of 0.02 M NaCI. Similarly, Cheng [Che85] also observed a 10% 
increase in adsorption of PEO with 0.02 M NaCI. Addition of NaCI up to 1 M 
strength to non-flocculating oxide suspensions, however, did not result in any 
measurable adsorption of PEO indicating that repulsion from a highly negative 
charged surface is not a major reason for non-adsorption of PEO. 
Effect of Hvdrated Counter-Ions 

The hypothesis of hydrated counter-ions preventing the approach of PEO 
molecules to the surface at higher pH assumes that the binding sites remain intact. 









54 
This implies that the adsorption of PEO at higher pH can be realized by contacting 
silica particles at a lower pH close to the isoelectric point and then increasing the 
pH. This hypothesis was experimentally observed at the solid/solution interface with 
the aid of AFM. 
AFM Studies 

Imaging of Adsorbed PEO Layer using Tapping Mode. The tapping mode 
of the AFM was used to image the adsorbed PEO molecules at i) pH 3.0 and (ii) pH 
9.5 by injection of 50 ppm of 5,000,000 MW PEO. In a separate experiment the 
initial pH was first maintained at 3.0 then changed to 9.5 and the desorption process 
observed. The AFM image of the bare silica plate is shown in Figure 4.12 with the 
average surface roughness determined to be less than 2 nm. 

The AFM image of the adsorbed polymer layer at pH 3.0 is presented in 
Figure 4.1 3. The thickness of the molecules was determined to vary from 25-40 nm 
which can be attributed to the polydisperse nature of the PEO molecules. The 
hydrodynamic thickness of adsorbed homopolymers on a saturated surface in a 
good solvent has been predicted to be between 2-3 R g [deG 81; deG 82; Sch82]. 
In the present case it thus seems that the probe did not detect the hydrodynamic 
thickness of the PEO molecules since the R g of the PEO molecule of 5,000,000 
MW is 144nm. 

Photon Correlation Spectroscopy (PCS) measurements for the latex-PEO 
system has shown that for PEO molecular weight greater than 280,000 the 
hydrodynamic thickness is more than 2 R g [Coh84;Cos84]. In the same system, 
Cosgrove et al [Cos84], however, found that Small Angle Neutron Scattering 



55 




8 

CQ 

1= 
D 
c/) 

n 

o 

55 



I— 

CO 

X! 
4- 

o 

<D 
OJ 

03 

E 






56 




Figure 4.13. AFM Tapping Mode topographic image of adsorbed PEO (MW= 
5,000.000 at pH 3.0) 



57 
(SANS) grossly underestimated the adsorbed layer thickness. The thickness of 
PEO molecule of 660,000 detected by SANS was found to be only 15 nm whereas 
with PCS a thickness of 95 nm was calculated. They attributed this discrepancy to 
detection of the segment density distribution of only the trains and loops and not of 
the tails by SANS. The segment density distribution of the adsorbed polymer has 
been shown to decrease exponentially with distance from the solid/solution interface 
[Cos84]. Thus a significantly lower thickness may be detected if the probe is not 
sensitive to the periphery of the adsorbed layer. 

The image of the adsorbed polymer molecules at pH 9.5 after an hour of pH 
change is shown in Figure 4.14. A much lower adsorption density of polymer 
molecules on the surface is observed. The image after another hour is similar to 
that of the virgin surface except a few isolated patches (see Figure 4.15). This, is 
the first reported direct proof of desorption of polymer molecules of high molecular 
weight upon pH change. 

It has been argued that at any given instant the probability that all the 
attached segments detach from the surface is so low that the adsorption for all 
practical purposes can be considered to be irreversible. However, de Gennes has 
pointed out the fallacy in this argument and showed that desorption is possible 
[deG87]. The desorption in the present study cannot be attributed to dilution since 
the polymer concentration was always maintained at 50 ppm. Further, the polymer 
was already adsorbed so that the possibility of inaccessibility to the surface due to 
the presence of hydrated counter-ions does not arises. The only reason which can 
then account for PEO desorption with increasing pH is the loss of binding sites for 
the ether oxygen of PEO. 



58 




:> T3 




— \ 




•c X 




\ c 




z 




ao 




o 




o o 




o • 




oo 




• o 




•^N 


0) 




03 


XN 




CO 




X 




Q_ 




+-* 




ns 




O 




LU 




Q. 




TJ 









-Q 




i— 




O 




W 




"O 


z 


03 


a, 


o 




8) 




O) 




03 




E 




o 




£ 




Q. 




03 




L. 




CD 




O 




a 




o 




-*— < 









T3 




O 




2 




D3 




C 




Q. 




Q. 




03 




H c 




5l 




< o 




to 




*3 




. *— 




■<* o 




2 3 




3 O 




D3.C 



59 




Figure 4.15. AFM Tapping Mode topographic image of adsorbed PEO at pH 9.5 
after two hours of desorption. 



60 
Force/Distance profiles with Contact Mode. The adsorption of PEO on the 
silica plate at pH 3.0 and its desorption at pH 9.5 was corroborated in the AFM 
studies by obtaining force/distance profiles using a glass sphere attached to the 
AFM cantilever. The results presented in Figure 4.14 indicate that at pH 3.0, in the 
absence of PEO, the net interaction profile does not show the presence of a 
repulsive force, as expected from the electrokinetic considerations and the DLVO 
theory. However, in the presence of PEO a steric repulsive force is observed 
revealing the adsorption of the polymer. On the other hand, at pH 9.5, the 
force/distance profiles remain virtually unchanged with and without the PEO 
indicating the absence of an adsorbed layer. 

The onset of the steric repulsion was observed at about 100 nm (see Figure 
4.16). Thus the thickness of the adsorbed layer is estimated be about 50 nm on 
each surface which is an underestimate with respect to the predicted and observed 
values in latex/PEO system [deG81; deG82; Sch82; Coh84; Cos84]. However, as 
mentioned above, the detection of the hydrodynamic thickness is dependent upon 
the sensitivity of the measuring probe to the peripheral layers of the adsorbed 
polymer. The measurement of force distance profiles to estimate the adsorbed 
layer thickness has been shown to be insensitive to the outer regions of the 
adsorbed polymer layer [Luc90]. Luckham and Klein [Luc90] measured the onset 
of repulsive interactions between adsorbed PEO layers (MW = 1,200,000) on mica 
in Surface Force Apparatus (SFA) at 190 nm. However, the thickness of the 
adsorbed PEO layer (95 nm) is about the R g (86 nm) of the PEO molecule. 
Further, these investigators have shown the effect of incubation time on the 
measured force/distance profiles. Measurements after only 1 hour of polymer 



61 



E 0.8 



a: 



0.6 




1*7 


pH 3.0 None 

X 


pH 3.0 PEO 


□ 


pH 9.5 None 


X 


pH 9.5 PEO 



20 40 60 80 100 120 
Separation Distance (nm) 






Fig 4.16. Effect of pH on interparticle forces between silica sphere and a flat plate 
with and without PEO(MW = 5,000,000). 



62 
contact time with the mica surfaces indicated the onset of interaction at 20 nm while 
after 16 h this value was found to be 190 nm. This behavior is not clear at present 
and further studies are needed to resolve this issue. An incubation time of 1 hour 
in the present case may, therefore, be responsible for a lower thickness of the 
adsorbed layer. 
Role of Specific Surface Binding Sites in PEO Adsorption on Silica 

It is clear from the AFM studies that there are specific PEO binding sites 
present on the silica plate at pH 3.0 which are lost when the pH is increased to 9.5. 
These are probably the most acidic silanol groups which dissociate at higher pH 
leading to the ionization of the site to a negatively charged species and loss of the 
bond with the ether oxygen. The ionization of the binding site to a negatively 
charged species is corroborated by the electrokinetic data (see Figure 4.4) and the 
force/distance profiles (see Figure 4.16). 

A similar mechanism was proposed by Evans and Napper [Eva73] to 
explain the decrease in PEO adsorption with increasing pH for the latex/PEO 
system. The binding sites on the latex were carboxylic acid groups which at low pH 
participated in hydrogen bonding with the ether oxygen of PEO. The ionization of 
the acid sites led to the loss of PEO adsorption at higher pH. 

It is also observed that the adsorption behavior of PEO as a function of pH 
on silica plate used in the AFM resembles that of silica B. The relative insensitivity 
of PEO adsorption on silica A to pH is in contrast to the behavior on silica B and 
silica plate and may be related to a different distribution of acidic sites and will be 
examined in Chapter 6. The lack of PEO adsorption on certain oxides which is not 
clear yet is discussed first in Chapter 05. 



CHAPTER 5 

ROLE OF SURFACE ACIDITY OF OXIDES IN PEO ADSORPTION 

Introduction 

In the last Chapter it was shown that no oxide other than silica exhibited 
significant adsorption of PEO. The mechanism of PEO adsorption on silica, 
consistent with the effect of pH, was indicated to involve the acidity of surface sites. 
In this Chapter this concept is discussed further to understand the reasons for lack 
of adsorption of PEO on other oxides. 

Koksal et al [Kok90] suggested that the lack of adsorption of PEO on alumina 
and hematite was due to the inaccessibility of the surface adsorption sites to PEO 
molecules. This explanation was found to be inconsistent with the effect of pH on 
PEO adsorption on silica in Chapter 4. The reason advanced by Koksal et al 
[Kok90] for the inaccessibility to the surface sites of PEO molecules, however, was 
not the hydrated counter ions but entropic factors which are discussed next. 
Accessibility of Surface Sites to PEO Molecules 

The net free energy change associated with the polymer adsorption must be 
negative and involves changes in both enthalpic and entropic factors. In the case 
of the oxides the processes contributing to the entropy changes on PEO adsorption 
are: i) the loss of water from the oxide surface, ii) loss of conformational entropy of 
the PEO molecule due to attachment of segments to the surface which were 
otherwise mobile and the iii) the entropy of dilution of the bulk phase. 



63 



64 
The ether oxygen itself has three water molecules attached to it and they 
have been shown not to detach in solution when PEO complexes with various ions 
[Pra95], The presence of the hydration layer has been demonstrated on Si0 2 , Ti0 2 , 
and Al 2 3 through measurement of hydration forces attributed to this layer [Gra93]. 
The disintegration of the hydration layer will be favored since it will result in more 
degrees of freedom for the water molecules along with dilution of bulk water. 
However, if the loss of conformational entropy of the polymer exceeds the gain in 
entropy by the loss of hydrated water molecules then the adsorption process is not 
favorable. In such a case, the surface sites will remain inaccessible to the polymer 
molecules. The replacement of water molecules on the surface is related to the 
concentration of the hydroxyl groups since these provide the adsorption sites for 
water molecules. 
Concentration of Surface Hydroxyls 

The concentration of surface hydroxyl groups on oxides has been determined 
by several investigators and is summarized in Table 5.1 [And82]. It is clear that 
alumina, hematite and titania are more hydrated than the silica surface and 
therefore release of water molecules is more favored from their surfaces. Thus the 
lack of PEO adsorption on alumina, hematite and titania does not seem to be due 
to entropic reasons. Koksal et al [Kok90], however, suggested that stronger 
hydrogen bonding of water molecules to the surface hydroxyls on hematite and 
alumina prevents the interaction of PEO with the latter. In such a case favorable 
entropy changes for PEO adsorption are expected to be affected the most for 



65 



Table 5.1. Concentration of surface hydroxyl groups [And82]. 



Oxide 



Silica 



Titania 



Hematite 



Alumina 



Number of OH/nm 2 



4.2-5.1 



4.9-6.2 



4.6-9.1 



15 



66 
hydrated oxides such as alumina and hematite. The argument of Koksal et al 
[Kok90] for stronger hydrogen bonding between the water molecules and surface 
hydroxyls was based on the number of surface hydroxyls per unit area and not on 
the energetics of water-surface hydroxyl interaction. Further, some overlap of the 
hydroxyl concentrations between the first three oxides indicates that the 
inaccessibility to the surface sites may not be the major reason for the non- 
adsorption of the PEO. 
Heat of Wetting of Oxides 

The enthalpy of interaction of water molecules with the surface hydroxyls is 
measured as the heat of wetting of the oxides (see Table 5.2)[Che59;Hea65]. It is 
observed that the heat of wetting of oxides follow the same trend as the 
concentration of surface hydroxyls. However, the interaction of water per surface 
hydroxyl show a considerable overlap for the oxides under consideration. This 
indicates that the water molecules are bonded with a similar strength to the surface 
hydroxyls on any oxide. The nature of surface hydroxyls, however, on different 
oxides may not be the same and this aspect is discussed next. 

Nature of Surface Hydroxyls 
Point of Zero Charge of Oxides 

For oxides with a hydrated surface, the surface chemistry in water is 
dominated by the chemical reactions 

MOH 2(surface) + = MOH surface + H solutlon + 

K,= [MOH surfa J [H solutlon + ] / [MOH 2(surface) + ] 

pK, = pH + log{[MOH 2(surface ;]/[MOH surfa J} 



67 



Table 5.2. Heat of Wetting values for oxides in water [Che59; Hea65]. 



Solid 


Heat of wetting 
(ergs/cm 2 ) 


Quartz 


260-370 


Amorphous silica 


165-220 


Rutile (TiO,) 


550, 550 


Fe,O a 


530 


ALO n 


650-900 



68 



M0H surface ~ M0 ' 1 surface + H+ solution 

K 2 = [M0- 1 surface ] [H + solutlon ]/[ MOH surfa J 
pK 2 = pH + log{[MOH surface ]/[MO" 1 surfa J} 
where M represents a metal ion at the surface. 

From these relations the point of zero charge (PZC) of the surface may be defined 
in terms of the pK's of reactions, i.e., 

PZC = 0.5 [pK,+ pK 2 ] 
and indicates the average acid-base characteristic of the surface. At any given pH, 
if an oxide surface donates relatively more protons than other it is more acidic and 
hence will have a lower pzc. In other words, the point of zero charge of an oxide is 
directly related to the acidity of the surface hydroxyl groups. The acidic surface 
hydroxyls are referred to as Bronsted acid sites (see Figure 5.1). 
Correlation between Heat of Wetting and pzc of Oxides 

Healy and Fuerstenau [Hea65] showed a linear correlation between the heat 
of wetting of oxides in water and their pzc (see Figure 5.2). This substantiates the 
assumption made in the preceding section that although the nature of the surface 
hydroxyls varies on different oxides the interaction strength with water molecules 
per surface hydroxyl is similar. Thus the role of solvent is not expected to be 
significantly different in PEO adsorption on the various oxides in aqueous medium. 



69 




W 






CD 






•*— i 






'w 






•a 






o 






03 






■a 






CD 






•4-* 






w 






c 




■ 


o 

1— 




C/) 

si 


CQ 




c 




i« 






«3 o 


$ 


c 


o 


o 


C Di 


.c 


a) o 




To 
o 


It 


03 


ll 


o -c 


E 


CD 


II II 


CD 


CL 


o >> 


-C 

o 

c/> 


3 


J 2 


Q. m CD 


T^ 






LT) 






CD 






i_ 






Z! 






O) 







70 



E 

1— 

CD 

ci) 

CD 



1000 



900 - 



800 - 



700 - 



600 - 



<: 500 - 



03 
CD 

X 



400 - 



300 - 



200 - 



100 





I 


I 


I 


I 


I I 


I 





Amorphous 


silica 




□ 

A 
V 


Quartz 
Tin oxide 
Hematite 








c 




O 




Titania 
Chromia 












O 


Alumina 






&/ 











- 








/ 


x s* 


- 


- 








/k 


- 


- 




q 

n r 


]/ 


n 






- 


l_l A. 




bJ 






/ C 


] 












I 


i 


I 


l 


I I 


i 



5 

PH 



8 



pzc 



Figure 5.2. Heat of wetting of oxides as a function of their point of zero charge 
(after [Hea65]). 



71 
Role of Bronsted Acidity in PEO Adsorption 
Strength of Bronsted Acid Site and Electronegativity of Cation 

It is shown above that the acidity of the Bronsted sites is the strongest on 
silica and the weakest on magnesia among the oxides under consideration. This 
can be understood qualitatively in terms of the electronegativity of the metal ion to 
which the surface hydroxyls are attached. The electronegativity of the surface metal 
atom governs the extent to which the electron pair shared between the metal and 
oxygen atoms is displaced towards the oxygen end. 

In case of a predominantly ionic bond such as in MgO the electron pair is 
close to the oxygen atom which results in a stronger attraction for the proton. 
Frequency shifts in the IR spectra observed during water and benzene adsorption 
on MgO indicated that these hydroxyl groups are more basic than those on silica 
surface [And65]. However, for an oxide with a significant character of covalent bond 
e.g. Si0 2 the proton will not be strongly bound to the oxygen and therefore this type 
of hydroxyl is expected to be acidic. Al 2 3 has a more covalent character of AI-0 
bond than Mg-0 but at the same time it is more ionic compared to Si0 2 . Thus, 
acidity of Bronsted sites is stronger on silica than on alumina. 

The dependence of electronegativity on the type of bonding and electronic 
environment was shown by Pauling [Pau63]. This implies that strict comparisons 
between oxides as regards to their surface acidity should be made only for a 
particular type. Thus for M0 2 type of oxide the pzc should increase in the order 

Si0 2 < Ti0 2 < Zr0 2 
because the electronegativity differences increase from 1.7 for silica to 2.0 for 
zirconia. The higher acidity of Bronsted sites on silica than titania thus explains the 



72 
adsorption behavior of PEO for the two oxides. This also illustrates the sensitivity 
of the interaction of the ether oxygen of PEO to the acid strength of the Bronsted 
sites. 
Relation Between Type of Oxide and its Point Of Zero Charge 

Parks [Par65] has summarized the broad probable ranges of the pzc 
characteristic of the cation oxidation state from the known literature values which 
are reproduced in Table 5.3. The relation between the pzc and the cationic size 
and charge was explained by an electrostatic model involving the coordination 
number with crystal field and hydration corrections. It is noted that the isoelectric 
points determined for the oxides under consideration are in close agreement with 
the predicted values given in Table 5.3. 

It is predicted from Table 5.3 that M0 3 and M 2 5 type of oxides should exhibit 
stronger Bronsted acidity than the other oxide types. The validation of Parks model 
[Par65] has been corroborated for Mo0 3 and V 2 5 by spectroscopic investigations 
using probe molecules of known acidity or basicity. 

It has been established through infra-red studies of adsorbed probe 
molecules that proton acid centers in simple oxides are essentially different in 
strength [Hai67;Dav90]. For instance, pyridine, a molecule only slightly less basic 
than ammonia, is not protonated on the surface of alumina which indicates that 
alumina has weak proton-donating properties [Cha63]. On the other hand, M0 3 and 
M 2 5 type of oxides such as MoO 3 and V p Respectively protonate not only pyridine 
but such weak bases as propene and ethene [Dav90]. Experimental data indicate 



73 



Table 5.3. Probable ranges of pzc of different types of oxides [Par65]. 



Oxide 


M 2 


MO 


M 2 3 


M0 2 


M0 3I M 2 5 


pzc, pH 


>11.5 


8.5-12.5 


6.5-10.4 


0-7.5 


<0.5 



74 
that the strongest Bronsted centers may be associated with the presence of Mo 6+ 
ions. Thus oxides such as Mo0 3 and Y Q are expected to exhibit stronger 
Bronsted acid sites than silica and should adsorb and flocculate with PEO. 
Adsorption and Flocculation Behavior of Mo0 3 and V 2 Q 5 with PEO 

Assuming that it is the presence of Bronsted acid sites with a strength 
comparable to or more than the acidic hydroxyls on surface of silica which is 
required to facilitate adsorption of PEO, Mo0 3 and V 2 5 should exhibit adsorption 
of PEO and flocculate. 
Electrokinetic studies 

It is observed from the electrokinetic behavior of Mo0 3 and V 2 5 shown in 
Figure 5.3 that both the materials exhibit high negative zeta potentials in the pH 
range studied. According to the prediction of pzc in Table 6.2, both the oxides have 
a pzc of less than pH 2.0. The electrokinetic studies on Mo0 3 and V 2 5 are also 
in agreement with the reported pzc values of 0.5 and 1 .5 respectively [Par65]. The 
suspensions for both the samples drifted back to the natural pH of about 2.7 within 
an hour of measuring the zeta potentials. Thus, adsorption and flocculation tests 
with these oxides were attempted only at pH 3.0. 
Adsorption of PEO on Mo0 3 and V z O z 

Adsorption isotherms for both Mo0 3 and V 2 5 with PEO of MW = 5,000,000 
are shown in Figure 5.4. Both the isotherms are of high affinity type similar to that 
for silica (see Figures 4.10 and 4.1 1). The saturation adsorption densities of all the 
oxides as a function of their acidic strength, which is measured by the pzc, are 
summarized in Figure 5.5. The saturation adsorption densities of PEO on all the 



75 







■10 



> 


-20 


E 








03 




-t— • 




c 
(1) 


-30 











0_ 




ro 




■*- 




(1) 




N 


-40 



-50 



-60 



T I 

O Mo0 3 
D V 2 5 



J L 



"i 1 1 r 




12 3 4 



5 
PH 



J L 



6 7 8 9 



Figure 5.3. Electrokinetic behavior of Mo0 3 and V 2 5 suspensions 
as function of pH. 



76 



1.0 



CM 



E 0.8 



■a 

CD 
.Q 

o 0.6 
w 

< 

CD 

E 

o 0.4 
a. 



c 

Z3 

o 
E 
< 



0.2 - 



0.0 







1 1 

o 


1 

O 


i 






' □ B 




a 


- 




1 UJ LJ 

/ / D 




a 




[ 


] 


o 


v 2 o 5 




[ 


] 


a 


Moo 3 




c 


) 








c 
1 


] 


i 


1 



20 40 60 80 

Residual Concentration, ppm 



100 



Figure 5.4. Adsorption isotherms of PEO on Mo0 3 and V 2 5 suspensions 
(PEO MW = 5,000,000; pH 3.0). 



77 




8 9 10 11 12 



PH 



pzc 



Figure 5.5. Saturation adsorption density of PEO (MW = 5,000,000) 
at pH 3.0 as a function of the point of zero charge of the oxides. 



78 
three oxides are similar indicating a similar adsorption mechanism. The insignificant 
adsorption of PEO on Ti0 2 indicates the sensitivity of the interaction of the ether 
oxygen with the surface Bronsted acid sites. Only those M0 2 type oxides with a pzc 
lesser than that of silica are expected to adsorb PEO. 
Flocculation behavior of Mo0 3 and V 2 Q 5 with PEO 

The flocculation behavior of Mo0 3 and V 2 5 as a function of the molecular 
weight of PEO is plotted in Figure 5.6. It is observed that for both the oxides the 
critical molecular weight for flocculation is 5,000.000 PEO. Large floes which were 
retained over 400 mesh screen, as in the case of silica, were formed. This, to our 
knowledge, is the first report of flocculation of any oxide other than silica with PEO. 

The flocculation of the two oxides with PEO as a function of dosage is plotted 
in Figure 5.7. It is observed that the breadth of flocculation decreases with increase 
in the molecular weight of the flocculant. The breadth of flocculation is larger for 
Mo0 3 and V 2 5 than Si0 2 with PEO of 8,000,000 MW (see Figures 4.2 and 5.7). 
Silica flocculation is reduced to 20% at 8 mg/g while the other two oxides flocculate 
to the extent of more than 60% at dosage of 1 5 mg/g. The critical molecular weight 
of PEO for flocculation of silica is 8,000,000 while that for Mo0 3 and V 2 5 is 
5,000,000. It appears from these observations that the breadth of flocculation 
increases with the flocculant molecular weight greater than the critical value. 
Surface charge and PEO adsorption 

It was shown in Chapter 4 that a negatively charged surface repelling PEO 
molecules is not a valid mechanism to explain the decrease of PEO adsorption with 
increasing pH. Further evidence that the PEO adsorption is not affected by the 



79 



100 



T3 
CD 

1 

=3 
O 
O 

o 



C 

=3 

o 
E 
< 




0e+0 2e+6 4e+6 6e+6 

Polymer Molecular Weight 



8e+6 



1e+7 



Figure 5.6. Flocculation behavior of Mo0 3 and V 2 O s as a function 
of PEO molecular weight (dosage = 0.5 mg/g; pH = 3.0). 



80 



CD 

o 

O 

_g 

LL 

C 
3 
O 

E 
< 



uu 




1 1 


1 1 


1 1 


1 


1 


80 


- 


jjkvf^ 




cN 







60 












" 


40 


- 




\ 






- 


20 


— 




■ 








- 




V 


o 


M0O3-8M PEO 








□ 


M0O3- 5M PEO 















A 


V 2 5 - 8M PEO 




1 — 1 


1 — 1 






V 


V 2 5 - 5M PEO 


1 1 


i_i 
1 


u 
1 




1 


1 


1 1 







4 6 8 10 

Flocculant Dosage, mg/g 



12 



14 



16 



Figure 5.7. Flocculation of Mo0 3 and V 2 5 as a function of PEO dosage 
at pH 3.0. 



81 
negatively charged surface has been provided by adsorption and flocculation of 
Mo0 3 and V 2 5 with PEO at pH 3.0 (> -40 mV zeta potential). 

It was also shown in Chapter 4 that the presence of adsorbed ions, which in 
solution state are capable of complexing with the ether oxygen, does not 
significantly affect PEO adsorption. It is, however, possible that complexation of the 
dissolved ions from Mo0 3 and V 2 5 with ether oxygen may occur in solution and 
subsequent precipitation of the complex on the surface may lead to PEO 
adsorption. The complexation of Mo ions with PEO in aqueous solution was shown 
by Vassilev et al [Vas86], Tests were therefore conducted to isolate the role of 
dissolved ions in the PEO adsorption process. 
Effect of dissolved ions 

Dissolution studies. Mo0 3 suspension was aged for 1 hour and the supernatant 
used to condition Al 2 3 and Ti0 2 powders to determine the adsorption of dissolved 
ions on the Al 2 3 and Ti0 2 surfaces respectively (see Table 6.4). It is observed 
that the maximum adsorption occurs on Al 2 3 while the least on SiO . This is 
expected since at pH 3.0 Al 2 3 is positively charged while Si0 2 is slightly negatively 
charged and the dissolved Mo is present as negatively charged species Mo0 4 2 ~ 
[Kun89]. The formation of crystalline Mo0 3 on the underlying substrate occurs only 
when more than 4.5 Molybdenum atoms/nm 2 of the surface are present [Kun89]. 
In the present case, monomeric ions are expected on Si0 2 (< 1 Molybdenum/nm 2 ) 
while heptameric species, Mo^" 6 , and an octahedrally coordinated polymeric 
surface species are present on Al 2 3 and Tip (1-4.5 Molybdenum/nm 2 ) 
[Oka88;Kun89]. 



82 



Table 5.4. Dissolution behavior of Mo0 3 and V 2 5 powders and the adsorption 
of dissolved ions on other oxides. 



pH = 3.0 

Solids loading 2g/100 ml 

Aging time = 3600 s 

Amount of Molybdenum dissolved = 330.6 ppm (16.5 mg/g solids) 

Amount of Vanadium dissolved = 275.1 ppm (13.7 mg/g solids) 



Oxide 


Residual in solution 

after 6h 

(ppm) 


Amount of Molybdenum 

adsorbed 

(mg/m 2 ) 


Number of 

Molybdenum ions 

per nm 2 of oxide 

surface 


Ti0 2 


280.5 


0.23 


1.4 


Al 2 3 


268.5 


0.44 


2.7 


Si0 2 


320.5 


0.16 


1.0 



83 
PEO adsorption and flocculation of Mo coated oxides. The non-flocculation 
behavior of the molybdenum coated alumina and titania was corroborated by 
adsorption studies wherein no measurable adsorption of 8,000,000 MW PEO was 
detected indicating that adsorbed ions do not cause adsorption of PEO. Thus it is 
shown that adsorbed ions do not cause adsorption of PEO. It is expected though 
that formation of crystalline Mo0 3 on the surface of a non-flocculating oxide may 
lead to PEO adsorption. In another study surface modification of Al 2 , by 
surfactant coating was found to result in PEO adsorption [Ram88]. However, PEO 
adsorption occurred only after the initiation of hemi-micellization of the surfactant 
on the alumina surface [Ram88]. This may be due to the interaction between the 
induced hydrophobic sites on alumina through hemi-micelle formation and the 
hydrophobic (CH 2 -CH 2 )- moiety of the PEO. Thus surface chemical modification 
through formation of either crystalline Mo0 3 /V 2 5 or hemi-micelles of oxides not 
amenable to flocculation with PEO present alternative routes to induce PEO 
adsorption. 

Role of Lewis Acid Sites 
It has been shown so far that the adsorption of PEO on oxides is sensitive 
to the acidity of the surface Bronsted sites. The other type of acid sites present on 
the oxide surfaces are Lewis sites which are exposed cations with unsaturated 
valence as illustrated schematically in Figure 5.8. It is known that the strongest 
Lewis acid sites (Al 3+ ions) are found on alumina surface [Dav90]. In principle, an 
acid-base interaction should be expected between the exposed Al 3+ ions and ether 
oxygen of PEO. The Al 3+ Lewis acid site on exposure to water does not convert to 



84 




Figure 5.8. Schematic showing a Lewis acid site (exposed cation). 
Large sphere = exposed cation 
Small spheres = oxygen atoms. 



85 
a Bronsted site [Parr63] and thus there exists a possibility of Al 3+ ion and ether 
oxygen interaction. 
Oxide/PEO/CCI, system 

van der Beek [Van91] has shown that in CCI 4 solvent the adsorption energies 
for various polymers are larger on silica than on alumina. The strength of the 
segment-surface interaction for different polymers varied in the same fashion for 
both alumina and silica. The adsorption energy of PEO was determined to be 1 kT 
higher on silica than on alumina and this was the largest energy difference between 
adsorption on silica and alumina for the polymers examined. This energy difference 
on silica and alumina may indicate that OH groups are more accessible for ether 
groups than Lewis acid sites. Ether groups in the main chain of adsorbate 
molecules are less exposed than functional end groups present on other polymers 
and are therefore more sterically hindered to acquire optimal orientations on the 
substrate. This steric hindrance will be more for Lewis acid sites than for hydroxyl 
groups because Lewis acid sites have more rigidly fixed positions on the surface 
than the protons taking part in hydrogen bonding. The hydroxyl groups have the 
ability to rotate and may therefore adjust their direction to the adsorbing groups for 
optimal interactions. Hence, the adsorption energy is significantly affected by steric 
hindrance between adsorbate molecules and Lewis acid sites compared to 
interaction between the ether groups and Bronsted acid sites. 
Hematite/Starch/Water System 

The interaction of Lewis acid sites with the adsorbate polymer molecules has 
recently been illustrated in the hematite-starch system [Pra91;Wei95]. Hydrogen 



86 
bonding between starch hydroxyl groups and mineral surface hydroxyl groups has 
been the favoured mechanism for many years mainly because of the presence of 
a large number of hydroxyl groups on the starch and mineral surfaces [Iwa82]. 
Later, evidence was obtained to suggest that a chemical interaction between the 
polysaccharide and the mineral is the likely mechanism for adsorption [Kho84]. 
Pradip postulated this interaction as a molecular recognition mechanism of Fe 
surface sites and the ether oxygen groups in starch molecule [Pra91]. Recently, 
Weissenhom and co-workers [Wei95] have shown through DRIFT studies that Fe 
sites on the surface of hematite participate in interaction with amylopectin and that 
hydrogen bonding plays a minor role in starch adsorption. 

An important conclusion from the hematite-starch system is that hydrogen 
bonding cannot be postulated as an interaction mechanism just because seemingly 
appropriate functional groups are present on the adsorbate and the adsorbent. It 
is the chemical nature of the surface sites that determines the bonding to the 
polymer functional groups. It must be noted that the functional groups in starch are 
not in the backbone chain like the ether oxygen in PEO and hence interaction with 
the Lewis acid sites is more probable for the former. The PEO molecule, however, 
also shows specificity of hydrogen bonding to oxide surfaces by interacting only with 
strong Bronsted sites as shown in the preceding sections. 



CHAPTER 6 
CHARACTERIZATION OF PEO BINDING SITES ON OXIDE SURFACES 

Introduction 

It is generally accepted that the hydroxyl coverage on oxide surfaces occurs 
as a result of water dissociation, assuming that every surface oxygen joins the 
hydrogen atom and the water OH groups are bound to metal atoms (see schematic 
in Fig 6.1). The evidence for chemical surface hydration of oxides has been 
provided by infrared absorption studies, heat of wetting measurements, and the 
thermal behavior of the adsorption-desorption kinetics of water. These studies have 
also shown that the surface hydroxyls on a given oxide are not equivalent in their 
chemical nature [You58;Hai67;Roc75]. 

It was observed that although silica A and silica B show similar affinity for 
PEO at pH 3.0, in Chapter 4, their behavior was significantly different at pH 9.5. 
Previous work on silica-PEO/PVA system identified the isolated silanols as the 
principal binding sites for the ether oxygen of PEO [Rub76;Che85;Kha87]. In this 
Chapter surface characterization of the silica samples as well as other oxides via 
DRIFT studies is described to further elucidate the role of isolated hydroxyls in PEO 
adsorption. Additionally, it was shown in Chapter 5 that the acidity of the surface 
hydroxyls on oxides determines the interaction with the ether oxygen of PEO. The 
surface acidity of the oxides is also characterized by DRIFT studies using pyridine 
as the probe molecule. 

87 



88 







-o-f 



o 
o 





CD 

c/3 

03 

03 
C 
03 



CO 
CD 
O 

03 
4- 

Ifl 

o 

03 

> 

c 
o 

c 
g 

ro 

S? 

2 
■o 
>^ 

sz 

8 
■E 

=5 
C/> 

c 

1 

-C 
CO 

o 

03 

E 

CD 
CO 



CD 

en 



89 
Surface Hydroxyls on Oxides 

The DRIFT spectra of the oxide samples in the hydroxyl region (3000-4000 
cm" 1 ) are shown in Figures 6.2 and 6.3. 
Silica A 

The DRIFT spectra of silica A (see Figure 6.2) shows the presence of 
isolated hydroxyl groups as a shoulder at 3747 cm" 1 . An isolated silanol is not 
hydrogen bonded since the minimum O-H-0 distance between neighboring silanols 
exceeds about 0.33 nm, equivalent to the normal van der Waals 0--0 contacts for 
non-bonded oxygen atoms. In accordance with Tsyganenko and Filimonov [Tsy72], 
only one type of isolated hydroxyl was observed on the silica surface. They 
demonstrated that the ultimate number of isolated hydroxyl groups possible on an 
oxide is one unit less than the oxygen coordination number which is two in the case 
of silica. 

The hydrogen bonding interactions give rise to other types of hydroxyls on 
the silica surface. The vicinal silanols are hydroxyl groups located on neighboring 
sites such that they may hydrogen bond with one another. In Figure 6.2 these are 
characterized by the band at 3649 cm* 1 . The broad band at 3470 cm" 1 is due to the 
adsorption of molecular water. A similar spectra for the Stober silica was obtained 
by Cheng [Che85] and Khadilkar [Kha87]. 
Alumina-A 

a-alumina is of corundum structure and has an oxygen coordination number 
of three. Thus, at the most, only two types of isolated hydroxyls can exist on the 
alumina surface which are observed at 3735 and 3700 cm" 1 (see Figure 6.2). 



90 




Ti0 2 3660 



i i i i i i 
3000 3200 3400 3600 3800 4000 



i, 



Wavenumber (cm" ) 



Figure 6.2. DRIFT spectra of oxides in the hydroxyl region. 



91 




Fe 2 3 



3472 3628 



_L 



J I L 



J 



3000 3200 3400 3600 3800 4000 



Wavenumber (cm" ) 



Figure 6.3 DRIFT spectra of silica B and hematite in the hydroxyl region. 



92 
Several bands due to hydrogen bonded hydroxyls in the range 3660 -3500 cm 1 were 
also observed. The presence of molecular water on the alumina surface is 
indicated by the band at 3490 cm". The characterization of the surface hydroxyls 
on oc-alumina is in agreement with that of Morterra et al [Mor76]. 
Magnesia 

Magnesia has the NaCI structure where oxygen atoms are octahedrally 
surrounded by six metal atoms. In principle, five types of hydroxyls are possible for 
such a lattice and their presence is dependent upon the number of M-0 bonds (1 , 
2 or 3 respectively) intersected by the cleavage plane. The bands attributed to 
isolated hydroxyls, as shown in Figure 6.2, occur at 3745 , 3700 and 3647 cm" 1 . 
Anderson et al [And65b] reported the presence of isolated hydroxyls at 3750 and 
3630 cm 1 for a MgO crystal where the cube face (100) was predominant. The band 
at 3512 cm" 1 can be ascribed to mutually bonded hydroxyls while that at 3447 cm 1 
is due to the presence of molecular water. At this stage it is important to note that 
the isolated hydroxy! groups for magnesia occur at a similar wavenumber as on 
silica. However, shifts observed during water and benzene adsorption on magnesia 
indicate that these hydroxyl groups are more basic than on silica [And65b]. 
Titania 

The anatase form of titania showed two bands at 3715 and 3660 cm" 1 which 
are due to isolated and bonded hydroxyls respectively (see Figure 6.2). These 
band assignments are in accordance with the previous work of Primet et al [Pri71]. 






93 
Mo0 3 and V 2 Q C 

The DRIFT spectra of Mo0 3 and V 2 5 as illustrated in Figure 6.2, is relatively 
featureless in the 3400-3800 cm" 1 region. This behavior has been attributed to the 
high mobility of the proton on the surface of these oxides [Dav90]. 
Hematite 

The DRIFT spectra of hematite presented in Figure 6.3 showed a broad band 
at 3130 cm" 1 due to adsorption of molecular water on the surface. The band 
positions of the isolated and hydrogen bonded hydroxyls are also indicated on the 
spectra [Hai67]. 
Silica B 

The surface of silica B shows the presence of hydrogen bonded hydroxyls 
and molecularly bonded water while the characteristic band of amorphous silica at 
3747 cm" 1 is not observed (see Figure 6.3). This may be due to the crystallinity of 
silica B which apparently presents a different surface structure than silica A which 
is completely amorphous. 

Isolated Hydroxyls and PEO Adsorption 
The band positions of the various types of hydroxyls along with the 
saturation adsorption density of 5,000,000 MW PEO are summarized in Table 6.1. 
The absence of isolated hydroxyls on silica B, Mo0 3 and \£ Q along with the 
presence of isolated hydroxyls on non-adsorbing oxides indicates that there is no 
correlation between the presence of isolated hydroxyls and adsorption of PEO. 
However, such a correlation may exist for isolated silanols and PEO adsorption 
[Che85]. 



94 



Table 6.1. Surface hydroxyls on different oxides and the saturation adsorption 
density of PEO. 






Oxide 


Hydroxyls 


Adsorption Density*, 
mg/m 2 


Isolated 


Bonded 


Mo0 3 


Not observed due to high mobility of 
proton. 


0.73 


V 2 5 


0.84 


Si0 2 (A) 


3747 


3649 
3470 


0.80 


Si0 2 (B) 




3470 


0.63 


Ti0 2 


3715 


3660 


<0.1 


Al 2 3 


3735 
3700 


3660 
3627 
3555 
3535 
3490 


Fe 2 3 


3628 


3472 
3130 


MgO 


3745 
3700 
3647 


3512 
3447 



ft Saturation adsorption density, measured at pH 3.0 with 5,000,000 MW PEO. 



95 
Rubio and Kitchener [Rub76] and Cheng [Che85] suggested isolated silanols 
to be the primary binding sites for PEO adsorption on the basis of changes in 
adsorption and the concentration of the isolated siianols with heat pretreatment. 
Upon heat treatment, the molecular water and hydrogen bonded hydroxyls are 
driven off the surface leading to an increase in the relative concentration of the 
isolated silanols. Thus heat treatment studies, therefore, were carried out for both 
types of silica samples to further elucidate the role of isolated silanols in the 
adsorption of PEO. Also, the accessibility of the surface sites to the PEO molecules 
was investigated by dehydroxylating the surface of alumina through heat 
pretreatment. 

Effect of Heat Pretreatment 
Silica 

A comparison of the DRIFT spectra for heat treated silica A with the 
untreated sample in Figure 6.4 reveals that the isolated hydroxyl peak for silica A 
sharpens with temperature and at 1 100°C only isolated silanols are present. 

The surface of silica B when heat treated to 800°C reveals an essentially 
dehydroxylated surface. The different speciation of OH groups on silica B may be 
attributed to its crystalline nature which presents a different surface structure than 
the amorphous silica A [Ile79]. The speciation of OH groups on amorphous silica 
surface has been shown to depend on the siloxane ring size, degree of ring 
opening, number of OH per surface silicon site [Bri90] and also surface curvature 
[Ile79] (see Figure 6.5). The convex surface, as of spherical silica A particle, may 
cause the neighboring OH to be isolated in nature (see Figure 6.5). 



96 



1100°C 




Silica A 



Silica B 



800°C 



I L. 



j L 



J L 



3000 3200 



3400 



3600 3800 



4000 



Wavenumber (cm ) 



Figure 6.4. DRIFT spectra of heat treated silica samples. 



97 







(i) 







H HH H HH 



b oooooo 




Figure 6.5. Schematic of i) amorphous silica surface showing the ring structure 
and ii) influence of surface curvature on H-bonding. A. Small positive radius of 
curvature reduces H-bonding between neighboring silanols. B. Large radius of 
curvature facilitates more hydrogen bonds. C. Negative radius of curvature 
exhibits strongest H-bonding. 



98 
Alumina-A 

In the case of alumina, heat treatment at 800°C resulted in the prominence 
of the bands corresponding to the isolated hydroxyl groups at 3735 and 3700 cm" 1 
(see Figure 6.6). The hydrogen bonded hydroxyls decreased and are not clearly 
resolved in the IR spectra. The band attributed to molecular water at 3490 cm" 1 
(see also Figure 6.2) shifts to lower wavenumber of about 3350 cm" 1 . At 1 100°C, 
the intensities of all these bands are significantly reduced. A similar evolution of IR 
spectra with heat treatment on Alumina A was earlier reported by Dow [Dow92]. 
Adsorption of PEO on Heat Treated Samples 
Silica 

It is seen from Figure 6.7 that silica A, which exhibits only isolated silanols 
at 1 100°C, adsorbs PEO confirming that isolated silanols are indeed the principal 
binding sites for PEO adsorption. The adsorption density of the heat treated 
surface (0.45 mg/m 2 ), however, was slightly lower than the untreated sample of 
silica A (0.62 mg/m 2 ). This follows the trend of saturation adsorption density of PEO 
and the concentration of isolated silanols on silica upon heat pretreatment [Che85]. 

Rubio and Kitchener [Rub76] postulated that the isolated hydroxyls on the 
silica samples disappeared when heated over 1000°C which led to insignificant 
adsorption of PEO. The existence of isolated silanols at 1 100°C and the adsorption 
of PEO on silica A, pretreated to 1100°C, appears to be in contradiction to their 
observations. This apparent discrepancy may probably due to the synthesis 
conditions along with high surface area (greater than 50 m 2 /g) of the silica samples 
used in their study. In the present investigation the silica A has a surface area of 



99 




1100°C 




800°C 



3700 



j 



3000 



3200 



3400 



3600 



3800 



4000 



Wavenumber (cm ) 



Figure 6.6. DRIFT spectra showing the effect of heat treatment on the surface hydroxylation 
of alumina A. 



100 



CM 



1.0 
0.9 
0.8 
? 0.7 



T3 
<D 

■2 0.6 
o 

GO 

< 

*- 5 

E 

I 0.4 

o 

§ 0.3 

E 
< 



0.2 



0.1 



0.0 




"<0" 



o 



O silica A -1100°C 



Q 



■e- 



o 



Silica B/ 800°C, Alumina A/800° ,1 1 00°C 

_J I I I I L 



i 







10 20 30 40 50 60 

Residual Concentration, ppm 



70 80 



Figure 6.7. Adsorption isotherms of PEO for heat-treated oxides at pH 9.5 
(PEO MW = 5,000,000). 



101 
about 3.2 m 2 /g. The viscous sintering, therefore, could have initiated at a lower 
temperature for high surface areas silica thereby leading to annihilation of the 
isolated silanols at lower temperatures. 

The absence of hydroxyls on silica B in the DRIFT spectra was reflected in 
the loss of adsorption of PEO compared to the untreated sample (see Figure 4.10). 
The lack of PEO adsorption on a dehydroxylated silica surface hass been attributed 
to a hydrophobic surface consisting of essentially siloxane groups [Rub76]. 
Alumina 

Heat treatment of Alumina A did not seem to change the low PEO saturation 
adsorption density as shown in Figure 6.7. The lack of adsorption of PEO on 
alumina irrespective of pH and heat pretreatment thus indicates that the surface 
hydroxyls are not acid enough to interact with the ether oxygen. Robinson et al 
[Rob64] found that the isoelectric point of alumina decreased to about 6.7 from 9.0 
upon calcination of the sample at 1400°C. Thus heat pretreatment alone does not 
appear to induce sufficient acidity of surface hydroxyls to facilitate interaction with 
the ether oxygen of PEO. 

The calcination of alumina leads to a progressively dehydroxylated surface 
which should mitigate the inaccessibility of the PEO molecules to the Lewis acid 
sites. Simultaneously, partial dehydroxylation is expected to result in an increase 
in the concentration of the Lewis acid sites. However, the absence of adsorption 
of PEO on partially dehydroxylated alumina further supports the concept that the 
strength of the Bronsted acid sites determines the binding of the ether oxygen of 
PEO. 



102 
Characterization of Surface Acidity of Oxides 

A measure of the strength of Bronsted acid sites is the point of zero charge 
of the oxide, as discussed in Chapter 05. Additionally, the surface acidity can be 
characterized by adsorption of probe molecules of known basicity. The pyridine 
molecule (k b =10 "*) is a weaker Lewis base than ammonia (k b =10 ') and is therefore 
more sensitive to the type of acid site [Par63]. The various possible complexes that 
form upon adsorption and the corresponding frequencies are shown in Table 6.2. 
DRIFT Spectra of Adsorbed Pyridine on Oxides 

Pyridine adsorption shown in Figure 6.8 revealed that Mo0 3 and V 2 5 exhibit 
sharp peaks corresponding to the formation of pyridinium ions at 1485 and 1535 
cm" 1 characteristics of the presence of strong Bronsted acid sites in agreement with 
previous studies [Dav90]. Silica A shows the presence of hydrogen bonded pyridine 
(bands at 1445 and 1595 cm" 1 ) indicating that the Bronsted acid sites are weaker 
than those present on Mo0 3 and V 2 5 . However, silica B, alumina A, magnesia, 
titania and hematite did not exhibit pyridine adsorption. The presence of stronger 
hydrogen bonding sites on silica A than alumina A indicated by pyridine adsorption 
is in agreement with the results of Parry [Par63]. It was determined that isolated 
silanols hydrogen bond with the pyridine molecules and can be removed by 
evacuation at 150°C [Par63]. On the other hand, hydrogen-bonded pyridine is 
removed from the alumina surface by evacuation at 25°C [Roc76]. The presence 
of weak Bronsted acid sites on alumina, magnesia, titania and hematite 



103 



Table 6.2 Infrared bands of pyridine in the 1400-1700 cm" 1 region of the spectrum. 
Band intensities are vs = very strong; s=strong; v=variable. 



Hydrogen Bonded 
Pyridine 


Coordinatively Bonded 
Pyridine 


Pyridinium Ion 


1440-1447 (vs) 


1447-1460 (vs) 


1485-1500 (vs) 


1485-1490 (w) 


1488-1503 (v) 


1540 (s) 


1580-1600 (s) 


1580 (v) 


1620(s) 




1600-1633 (v) 


1640 (s) 



1484 



1534 




V 2°5 




MoO, 



104 




1445 



1595 



1400 



1500 



1600 



Wavenumber (cm ) 



1700 



Figure 6.8. DRIFT spectra of pyridine treated M0O3, V 2 5 and Si0 2 
samples. 



105 
incapable of hydrogen bonding with pyridine, is in agreement with the 
characterization of surface acidity by point of zero charge of oxides. 
Acidity of silanol groups 

The absence of hydrogen bonding with pyridine for silica B is in contrast to 
silica A and may be due to a significantly lower concentration of strong Bronsted 
acid capable of bonding with pyridine. Model studies on silsequioxanes have 
revealed that sites of differing reactivity may be present depending on the extent of 
hydrogen bonding between the neighboring hydroxyls on the silica surface 
[Feh89;Feh90]. Isolated silanols were predicted to be less acidic than clusters 
possessing at least three mutually hydrogen bonded hydroxyl groups. This was 
experimentally shown to be the case by the calorimetric adsorption results of 
Chronister and Drago [Chr93]. These authors elucidated three hydrogen-bonding 
sites of different strengths with the isolated silanols being the weakest acidic site on 

the surface. 

Surface Analysis of Silica and Adsorption of PEO 
Specificity of hydrogen-bonding of isolated silanols 

The isolated silanol sites, although not the most acidic, are still the stable 
sites for PEO adsorption as illustrated by the relative insensitivity of the adsorption 
isotherms for silica A to pH (see Figures 4.9 and 4.10). This stability may be due 
to their large concentration on the silica A surface which is exemplified by a 
significant adsorption of PEO on silica sample pretreated to 1100°C (see Figure 
6.7). As a result of this pretreatment the isolated silanols were the only adsorption 
sites available on the silica surface. 



106 
In accordance with the results of Cheng [Che85] and Behl and Moudgil 
[Beh93d], no shift in the frequency of the isolated silanol groups on PEO coated 
silica A surface was noticed to indicate the specificity of the hydrogen bonding. This 
has been attributed to the presence of water molecules [Che85;Beh93d]. The effect 
of water molecules on the sensitivity of the measurements can be obviated by the 
use of non-aqueous solvents. In non-aqueous solvents direct evidence for 
hydrogen bonding of the isolated silanols to a copolymer of ethylene oxide-methyl 
methacrylate was obtained where large shift (about 100 cm" 1 ) was observed for the 
polymer coated silica particles [Fri62;How70]. 

The specific nature of the hydrogen bonding interactions of the isolated 
silanols with several organic molecules in vapor state including diethyl ether (the 
monomer group of PEO) was established by Kiselev [Kis65]. He determined that 
the specific heat of adsorption on the hydroxyl group, Q a is approximately linearly 
related to the frequency shifts v 0H obtained upon adsorption of these molecules. 
This relationship is reproduced in Figure 6.9 [Kis65]. Basila [Bas61] proposed that 
since the heat of adsorption is inversely proportional to the ionization potential of the 
adsorbate, this relationship serves as evidence that the interaction between the 
hydroxyl groups and the adsorbate is essentially one of charge transfer. It is 
observed that the ether oxygen produced a larger shift (450 cm" 1 ) than water (200 
cm" 1 ) indicating a stronger hydrogen bonding of the isolated silanols to the former 
[And65a;Kis65]. 



107 



500 



400 



300 



E 

i 
o 



200 



100 







i r 



n-C 6 H 14 



i 1 1 r - 

CH2CH 2 0CH 2 CH2 

(C 2 H 5 ) 2 ( 



CH 2 CH 2 COCH 2 CH 2 



CH 3 CN 




CH 3 N0 2 



CH 3 COOC 2 H 5 



J I I L 



1 



2 3 4 

Q (kcal/mole) 



Figure 6.9. Plot of the change in frequency of isolated silanols against 
the specific heat of adsorption for several vapors adsorbed on silica 
surface (data after Kieslev [Kis65] and Anderson [And65a]). 



108 

Effect of pH 

The adsorption of PEO at pH 3.0 on silica B (see Figure 4.10) and the 
absence of isolated silanols indicate that other types of silanols with sufficient acidity 
to hydrogen bond with the ether oxygen of PEO must be present on the surface. 
However, at pH 9.5 a significantly lower PEO adsorption was determined on silica 
B and the silica plate. This suggests that with increasing pH the concentration of 
the binding silanol sites is considerably reduced. These silanols interact with the 
ether oxygen close to the isoelectric point, where they are present in the unionized 
form. Their ionization to a negatively charged surface species with increasing pH 
results in the loss of binding to the ether oxygen of PEO and a corresponding loss 
in PEO adsorption. 

The presence of hydrogen bonded silanols on the silica surface, which are 
more acidic than the isolated silanols, has been predicted and experimentally 
validated, as discussed in the preceding section [Feh89;Feh90;Chr93]. Their 
dissociation will precede that of the isolated silanols on account of their higher 
acidity. The lower acidity of the isolated silanols along with their large 
concentration, estimated at about one-third of the total surface silanols [Arm69], 
may then be responsible for the relative insensitivity of PEO adsorption on silica A 
to pH changes. 



CHAPTER 7 

ADSORPTION AND FLOCCULATION BEHAVIOR OF SILICATES WITH PEO 

Introduction 

The silicate minerals essentially comprise of Si0 4 tetrahedra and oxygen 
or hydroxyls octahedra containing divalent or trivalent cations. The different type 
of silicate structures, as discussed in Chapter 3, are essentially classified by the 
connectivity of the Si0 4 tetrahedra. Thus the building blocks of silica and other 
oxides are present in the silicates. 

The adsorption of PEO on silica and its flocculation behavior has been 
discussed in the rest of this study. The other simple oxides with a pzc greater than 
that of silica did not exhibit adsorption of PEO. Thus it is of interest to examine the 
adsorption behavior of PEO on mixed oxides containing silica and other oxides in 
the context of the PEO adsorption mechanism determined for oxides. In this 
Chapter the adsorption behavior of PEO on representative samples of each silicate 
type is discussed and the adsorption mechanism of PEO correlated with their 
structural characteristics. 

Adsorption and Flocculation Studies 
Flocculation of silicates 

The flocculation behavior of the silicate samples as a function of the 
molecular weight of PEO is shown in Figure 7.1. It is observed that only layered 



109 



110 



a 
o 
o 

LL 

is 
o 

E 
< 



100 



80 



60 



40 



20 



00 



O Palygorskite 

□ Kaolinite 

A Tremolite 

V Augite 




100000 



1000000 
Molecular Weight 



10000000 



Figure 7.1 . Flocculation of silicate samples as a function of molecular 
weight of PEO (pH = 9.5; dosage = 1mg/g) 



111 

(kaolinite and palygorskite) and chain (augite and tremolite) silicates flocculate while 
the orthosilicate minerals (almandine, olivine and topaz) do not. The critical 
molecular weight of PEO for flocculation of the silicate samples is summarized in 
Table 7.1. It is seen that for the clays and tremolite the critical molecular weight is 
900,000 while for augite it is 4,000,000. It is well known that palygorskite, a 
predominant component of phosphatic clays, flocculates easily with PEO 
[Sch86;Bro89]. Similarly the flocculation of kaolinite with PEO was shown by Koksal 
et al [Kok90]. The flocculation behavior of pyroxene and amphibole type of silicates 
has been reported in the literature. The flocculation behavior of these two silicate 
minerals, therefore, is examined in detail in the following section. 
Effect of Flocculant Dosage 
Augite 

It is observed that the flocculation of augite is independent of the dosage of 
PEO of 8,000,000 MW whereas a maximum in flocculation at a dosage of 0.5 mg/g 
is seen with PEO of 5,000,000 MW (see Figure 7.2). In fact, the flocculation 
behavior of augite with PEO is similar to that of Mo0 3 (see Figure 5.7). The 
optimum dosage of 5,000,000 and 8,000,000 MW PEO is similar for both the 
materials, being about 0.5 mg/g and 0.1 mg/g respectively. This implies that the 
number of molecules of 8000,000 MW PEO adsorbed per particle at the optimum 
dosage is nearly an order of magnitude lower than that for 5,000,000 MW PEO. 
This indicates that bigger polymer molecules form more or stronger bridges due to 
more segments in contact with the particle surface, although no quantitiative 
explanation for this effect has been proposed yet. A similar result was obtained for 



112 



Table 7.1. Critical molecular weight for flocculation of the silicate minerals at pH 
9.5. 



Material 


Critical Molecular 
Weight of PEO 


Palygorskite 


900,000 


Kaolinite 


900,000 


Tremolite 


900,000 


Augite 


4,000,000 


Almandite 


No flocculation 


Olivine 


Topaz 



113 



100 
90 
80 
70 



Cl) 
05 


60 


13 




O 




O 

o 


60 


M— 




+-> 




c 

3 


40 


o 




E 




< 


30 




20 




10 








l 


l 1 1 l I 


- 


~~Q~~ § 




O PEO MW = 5,000,000 


- 


□ PEO MW = 8,000,000 


- 


I I I I 

QO 


i 


i i i i i 



2 4 6 8 

Polymer Dosage (mg/g) 



10 



12 



Figure 7.2. Flocculation behavior of augite as a function of PEO 
dosage at pH 9.5. 



114 
the latex-PEO system by Perssels et al [Per90]. As a result of fewer number of 
molecules of higher molecular weight polymer required for flocculation it may be 
possible that the flocculation rate is enhanced with respect to the adsorption 
kinetics. In such a case, prior to complete coverage of the surface by polymer 
molecules at higher dosages, stable floes may already have been formed. This 
may, thus, explain the relative insensitivity of flocculation of augite and Mo0 3 to the 
dosage of PEO of 8,000,000 MW. 
Tremolite 

The flocculation of tremolite as a function of PEO dosage is plotted in Figure 
7.4. It is observed that, similar to augite, the flocculation of tremolite is relatively 
insensitive to the polymer dosage for PEO molecular weight greater than the critical 
MW of 900,000. 
Adsorption Studies on Silicates 

The adsorption kinetics of PEO of 5,000,000 MW at a dosage 10 mg/g on 
tremolite and augite at pH 9.5 is presented in Figure 7.5. It is observed that after 
about 12 hours the adsorption reaches an equilibrium value for both tremolite and 
augite. An equilibration time of 16 hours, however, was used for deriving adsorption 
isotherms of PEO on the silicate minerals. 

The adsorption isotherms of PEO on tremolite, augite, almandite, olivine and 
topaz at pH 9.5 are plotted in Figure 7.6. The adsorption isotherms for the layered 
silicates, kaolinite and palygorskite are presented in Figure 7.7. It is observed that 
the orthosilicates exhibit much lower adsorption compared to the layered and chain 
silicates. Thus the equilibrium time of 16 hours for the orthosilicates is justified on 



115 



T3 
CD 

JS 

=3 

o 
o 
o 



c 

13 

o 
E 
< 



100 



80 



60 



40 



20 








O PEO MW = 900,000 

□ PEO MW = 4,000,000 

A PEO MW = 5,000,000 

V PEO MW = 8,000,000 



4 6 

Polymer Dosage (mg/g) 



8 



10 



Figure 7.3. Flocculation behavior of tremolite as a function of PEO dosage 
at pH 9.5. 



116 



7 



£ 
■a 

CD 

-Q 
i— 

O 
w 

< 

CD 

E 
j^ 

o 

0_ 



s 2 

o 
E 
< 

1 







-O o 




-D- 



_n_ 



XT 



TT 



-3 



o Tremolite 
□ Augite 



5 10 15 20 25 30 35 

Contact time (h) 

Figure 7.4. Kinetics of PEO adsorption on tremolite and augite. 



117 









10 



"a 
o 

< 



1 

O 

a. 



9 - 



£ 



7 



6 - 



5 - 



4 =; 



V 3 - 



c 

=3 
O 

E 
< 







1 


! 1 


1 1 





o 

a 


1 1 1 

Tremolite 
Augite 


- 










A 
V 


Almandite 
Olivine 




- 


QO^- 









Topaz 


- 


Q — o 


<e> 




O 


_ yr 


o 










- 


- U^- 








r-l 


□ 


- 




a 


!_1 


□ 


/u 












- 


\ 




« o. 




o 


V 


<X 


8 , 


\ 1 






1 


i i i 


z\ 







10 20 30 40 50 60 70 80 90 100 
Residual Polymer Concentration, ppm 



Figure 7.5. Adsorption isotherms of PEO on chain and orthosilicates 
at pH 9.5 (PEO MW = 5,000,000) 



118 



35 



30 

E 

T3 

CD 

-Q 

i— 

O 
w 

"O 
< 

L. 

CD 

E 
£ 15 



c 

3 
O 

E 
< 



25 



20 




10 



i r 








<0 



O Palygorskite 
□ Kaolinite 



15 30 45 60 75 90 105 120 135 150 
Residual Polymer Concentration, ppm 

Figure 7.6. Adsorption isotherms of PEO on clays at pH 9.5 
(PEO MW = 5,000,000). 



119 
the basis of the low PEO adsorption on these silicates. The adsorption density is 
normalized with respect to the weight of the solids since the layered and chain 
silicates are porous (see Table 3.6). The saturation adsorption density of PEO on 
the silicates is summarized in Table 7.2. The use of the BET surface area values 
is thus seen to underestimate the adsorption density on palygorskite, kaolinite, 
tremolite and augite. Estimation of the effective surface area, however, is required 
for comparing adsorption on samples of different porosity and is discussed next. 
Estimation of the effective surface area 

BET method. The hydrodynamic diameter of the molecules of PEO of 
5,000,000 MW varies between 91 nm and 225 nm (see Figure 3.2). Therefore, the 
effective surface area for polymer adsorption is not the total surface area 
determined from the BET measurements since pores below 91 nm size would be 
excluded in the adsorption process. Further, the effective surface areas 
corresponding to these pore diameters will give minimum and maximum values of 
the saturation adsorption density respectively. The BET surface area for tremolite 
for pore sizes of 91 nm or higher and 225 nm or higher is 0.297 and 0.023 m 2 /g 
respectively while for augite and the orthosilicates there is no corresponding 
detectable surface area. Thus the insensitivity of the BET method to large pore 
sizes renders it unsuitable for determining effective surface area for higher 
molecular weight polymers. 

Particle size measurements. The geometric surface area calculated from the 
particle size distribution of the tremolite and augite is about 0.24 m 2 /g. This area is 
calculated based on the assumption of equivalent spherical particles. 



120 



Table 7.2. Saturation adsorption density of PEO (MW = 5,000,000) on chain and 
orthosilicates. 



Material 


Flocculation 


Saturation Adsorption 
Density, mg/g 


Palygorskite 


Yes 


32.0 


Kaolinite 


Yes 


6.0 


Tremolite 


Yes 


5.2 


Augite 


Yes 


3.4 


Almandine 


No 


0.4 


Olivine 


No 


0.6 


Topaz 


No 


0.6 



121 

The saturation adsorption density of PEO, on this basis, is about 10 and 20 mg/m 2 
respectively. These values are an order of magnitude more than the saturation 
adsorption density of PEO on oxides which exhibited flocculation (see Figure 5.5). 

An estimate of the effective surface area of palygorskite was obained by 
considering its needle like morphology with the particles typically of 1.0 urn length 
and 0.05 urn in cross-section [Hog85]. The microporosity of the particles, which 
comprises of the intralayer channels for an individual crystallite and the interlayer 
channels between the crystallites, is inaccessible to PEO of 5,000,000 MW. Using 
the known density of palygorskite (2240 kg/m 3 ) an external surface area (the 
geometric surface area) of about 36 m 2 /g is obtained. This effective surface area 
corresponds to a saturation adsorption density of about 0.9 mg/m 2 and is similar 
to those determined for the flocculating oxides viz. Si0 2 , Mo0 3 and V 2 5 . 

Hg-porosimetrv. More realistic values of the surface area involved in the 
adsorption process for tremolite and augite were obtained using Hg-porosimetry 
(see Table 7.3). Also the saturation adsorption density of PEO, calculated on this 
basis is reported in Table 7.3. These values are reasonable considering that the 
saturation adsorption density of PEO of 100,000 MW was determined to be about 
4+1.5 mg/m 2 on mica, a sheet silicate [Kle84; Luc90]. 

In view of the above discussion, even though uncertainties are involved in 
estimation of the actual surface area involved in the adsorption process, a trend of 
a higher adsorption of PEO on the chain and layer silicates than the orthosilicates 
is observed. The higher adsorption density of PEO on the layered and chain 






122 



Table 7.3. Estimated surface areas from Hg-porosimetry and the calculated 
saturation adsorption desnsities for tremolite and augite. 



Material 


Effective surface area, 
m 2 /g 


Saturation Adsorption 
Density, mg/m 2 


Pore size 
(91 nm) 


Pore size 
(225nm) 


Pore size 
(91 nm) 


Pore size 
(225nm) 


Tremolite 


1.42 


1.00 


3.66 


5.20 


Augite 


1.38 


0.88 


2.46 


3.86 



123 
silicates than the oxides may be attributed to a different conformation of the PEO 
molecule on these substrates. AFM studies were conducted on polished samples 
of these minerals to gain further insight into the conformation of the PEO molecules 
on chain silicates. 

AFM Studies of Adsorbed Molecules on Tremolite and Auaite 
Tremolite 

The AFM image of the bare tremolite surface is shown in Figure 7.8a. The 
mean surface roughness was determined to be 0.2 nm while the maximum peak to 
valley difference is 1.0 nm. The image of the adsorbed PEO (MW = 5,000,000) 
molecules at pH 9.5 after 1 hour of adsorption is shown in Figure 7.8b. The 
presence of the polymer on the tremolite surface was corroborated by obtaining 
images in the friction mode where the adsorbed polymer molecules are 
characterized by regions of lower friction (see Figure 7.8c) [Cul94]. The histogram 
of the parking areas of the PEO molecules on tremolite is shown in Figure 7.8d. 
The interpretation of these images to explain the polymer conformation at the solid- 
solution interface is given below. 

The polymer concentration of 5 ppm was used which corresponds to 
achieving saturation adsorption density from the results obtained with the solution 
depletion technique. Although a complete coverage of the surface with the PEO 
molecules was expected the polymer was adsorbed as a spotty coating, i.e., the 
surface appears to be unsaturated (see Figures 7.b and c). This observation is not 
understood at present but significantly different factors from the adsorption 
experiments using the solution depletion technique such as the lack of turbulence 



124 




:> :> 
-d-d 

Z X 

3. C 

oo 
oo 

■ ■ 

OO 



XN 



8 

I? 

J 

O 

E 
8> 



X! 

"5 

<u 

03 
E 



3 
D3 



125 




N. I 

z c 
a. 

o 

oo 

oo 

in • 

■ o 



XN 



o 

E 

<D 

L. 

c 
o 

O 

UJ 
CI- 
TS 
(1) 
_Q 

o 

CO 
T3 
03 

M— 

o 

CD 
CD 
03 

E 

< 



oo 

=5 

en 



126 




| 

o 

E 

<D 

■•-» 
C 

o 
O 

LU 
Q. 



o 
w 

T3 

re 

o 
a> 

O) 

03 

E 

c 
g 

g 

i_ 

LL- 



CD 

u 

3 
O) 



127 










HistograM 







5000 A 10000 
Area CnM 2 1 



Figure 7.10. Histogram of parking area of PEO molecules on tremolite. 



128 
and a flat geometry of the tremolite surface in the AFM may have contributed to this 
behavior. 

The thickness of the adsorbed PEO molecules, as determined from section 
analysis of Figure 7.8b, is between 2.5-5.0 nm. The maximum and minimum values 
of the parking area of the PEO molecules, as obtained from Figure 7.8d, are 16,000 
and 770 nm 2 respectively and may be due to the polydispersity of the polymer 
sample. The histogram also revealed that the large molecules are fewer in number 
than the small molecules. The range of parking areas obtained from the histogram 
in Figure 7.8d correspond to PEO MW between 300,000 and 4,000,000 from the 
consideration of a flat conformation, i.e., with all the segments in contact with the 
surface. The molecular weight distribution of the PEO molecules of 5,000,000 MW 
in solution, as determined by light scattering, lies in a similar range. Thus the 
conformation of the adsorbed molecules appears to be flat which is in agreement 
with the theoretical predictions [Coh84;Cos90]. According to their theory the loops 
and tails develop only after adsorption of a monolayer of polymer molecules on the 
substrate. The optimum flocculation dosage is, however, always below the 
monolayer coverage and the occurrence of a nearly flat conformation seems in 
apparent contradiction to the predicted and observed conformation. However, the 
two seemingly opposite observations are explained by considering the dynamics of 
polymer conformation [Pel90]. The initial attachment of the polymer molecule to the 
surface is associated with a conformation similar to the random coil in solution and 
the polymer chain eventually relaxes towards a flat conformation (the equilibrium 
conformation) as long as the surface is unsaturated. The time scale of the 



129 
relaxation process has been estimated to be the order of a few seconds which 
agrees with the experimental observation that the flocculation is instantaneous. In 
the present AFM results the experimental observation of the adsorbed polymer is 
made after an hour during which the polymer molecules would have acquired a flat 
conformation. 

The thickness of the adsorbed PEO molecule on tremolite, from Figure 7.8b 
indicates a flat conformation. However, as observed from Figure 4.13c the 
thickness of the same polymer molecules on silica is between 25-50 nm even 
though the parking areas are similar. Thus the lower thickness could be due to the 
type of technique used to probe the adsorbed molecule in the AFM. In the contact 
mode, used for tremolite, the applied force is about two orders of magnitude more 
than in the tapping mode used for silica, and hence the contact mode is sensitive 
only to rigid materials. It is known that the segment density distribution of the 
adsorbed macromolecules falls of exponentially into the solution from the 
solid/solution interface [Coh84;Cos84]. In such a case, only the incompressible 
adsorbed polymer molecule, corresponding to the first few layers will be imaged in 
the contact mode. Further, the lower saturation adsorption density on silica (0.63 
mg/m 2 ) than tremolite (3.66 mg/m 2 ) precludes a flatter adsorption on tremolite than 
silica. Also, no evidence for multilayer or copious adsorption leading to high 
adsorption density on tremolite was observed. In this regard, further investigations 
are warranted to correlate the adsorption studies on a flat plate in the AFM with the 
solution depletion technique used for adsorption density measurements on 
powders. 



130 

Augite 

The mean roughness of the bare surface of the augite sample shown in 
Figure 7.9a was determined to be 0.3 nm and the maximum peak to valley 
difference was found to be 0.8 nm. The image of the adsorbed PEO molecules 
after 1 hour of adsorption at pH 9.5 is presented in Figure 7.9b. The presence of 
the adsorbed molecules was also corroborated by obtaining images in the friction 
mode similar to the case of tremolite (see Figure 7.9c). In contrast to tremolite, a 
larger number of PEO molecules appear to adsorb on the augite surface, though 
a complete surface coverage is not observed. The thickness of the incompressible 
layer of adsorbed polymer (less than 2.5 nm) is observed to be lower than on 
tremolite. The parking area of the PEO molecules, as indicated from the histogram 
in Figure 7.9d, was determined to be between 200 and 2,550 nm 2 corresponding to 
a size distribution between 15-70 nm. These values when compared with those 
obtained for tremolite indicate a more coiled molecule at the solid/solution interface 
in the case of augite, and thereby the important role of the surface characteristics 
in determining the conformation of the adsorbed polymer molecule. 

Conceptually, the adsorbed polymer in the form of a random coil invokes the 
picture of loops and tails in addition to the train segments. The involvement of 
segments in the formation of loops and tails is expected to result in a decrease in 
the parking area from a flat conformation of the adsorbed polymer molecule. 

An attempt was made to calculate the parking area of adsorbed PEO 
molecule with a random coil conformation following the model of Tronel-Peyroz 
[Tro83]. This model was used by Behl et al [Beh93c] to calculate the parking area 



131 




z z 

a. c 

oo 
oo 



XN 



8 

3 

3 
03 

ro 
-Q 



03 

E 

LL 

< 



ii. 



132 




"«\ 
*>w Z 

z c 
a. 

o 

oo 

oo 

lo • 

■ o 



XN 



O) 

3 
03 

C 

o 
O 

LU 
Q- 

T3 

I 

O 

w 
■o 

(0 

o 
<u 

O) 
(0 

E 



CM 



133 



Mmm^mmm 







03 

C 

o 
O 

LU 
Q_ 

TJ 

CD 

-e 

o 

(0 

•s 

0) 
OJ 

03 

E 

c 
o 

■8 



U- 

< 

CO 



>c 



in in 



134 



Grain size histograM 




o A 



15187 



Area CnM 2 ] 



Figure 7.14. Histogram of parking area of PEO molecules on augite. 



135 
of the adsorbed PEO molecule in the molecular weight range of 3,400-5,000,000. 
The calculated parking area, however, predicted a decrease in the saturation 
adsorption density with increasing molecular weight. The experimental results of 
Behl et al [Beh93c] for apatite-PEO and dolomite-PEO systems and also of Cohen- 
Stuart et al [Coh84] for latex-PEO system, however, show an increase in the 
saturation adsorption density with increasing molecular weight. Their results are in 
agreement with the predicted effect of molecular weight on saturation adsorption 
density [Sat80, Daw82. Lip74]. The Tronel-Peyroz model, in addition to predicting 
a reverse trend of the variation of saturation adsorption density with molecular 
weight, also calculates a higher parking area than that obtained with the assumption 
of a flat conformation for PEO of molecular weight higher than 600,000. Thus there 
is still a critical need for a suitable model for calculating the parking area of 
adsorbed polymer molecule with a random coil conformation. 

The AFM studies showed that the PEO molecules adsorbed at certain sites 
on the surface even though the initial polymer concentration was high enough to 
cover the whole surface. This may be an experimental evidence of the concept of 
active sites proposed by Behl et al [Beh93c] for polymer adsorption since due to 
surface heterogenities not all of the surface may be active for polymer adsorption. 
The chemical binding entities on the surface which comprise the region of an active 
site have been identified as highly acidic surface hydroxyl groups (Bronsted acid 
sites) (Chapters 4-6). Further, for silica, the isolated silanols were determined to 
be the PEO adsorption sites irrespective of pH. The adsorption and flocculation of 
silicates at pH 9.5 indicates the presence of such groups on silicate surfaces as 



136 
well. The surface characterization of the silicates was conducted to elucidate the 
presence of isolated silanols and other Bronsted acid sites for PEO adsorption. 
Surface Characterization of Silicate Minerals 

The DRIFT spectra for the chain and layered silicates are summarized in 
Figure 7. 1 while those for the orthosilicates are summarized in Figure 7.11. The 
observed positions of the bands in the hydroxyl region are in agreement with those 
reported by van der Marel and Beutelspacher [van76] 
Kaolinite 

The characteristic bands at 3696 and 3617 cm" 1 of the structural hydroxyls 
groups were observed in the DRIFT spectra shown in Figure 7.10. The band at 
3696 cm" 1 has been attributed to the hydroxyl groups belonging to the octahedral 
layer but opposite to the tetrahedral oxygens of the adjacent silica layer [Far64]. 
The band at 3617 cm" 1 is due to the hydroxyl groups between the tetrahedral and 
the octahedral sheets [Far64]. The weaker bands at 3650 and 3670 cm' 1 have 
been attributed to the outer surface hydroxyls - the conventional surface hydroxyls- 
which occur at the broken edges and on the octahedral layer that is exposed to the 
surface [Hai67]. The hydroxyls responsible for the band at 3696 cm" 1 have been 
considered to be isolated in nature [Grim68]. 
Palygorskite 

The band at 3619 cm 1 has been attributed to the isolated hydroxyls by Serna 
et al [Ser78]. These hydroxyls, however, are not the conventional surface hydroxyls 
but belong to the structure of palygorskite similar to kaolinite. The water molecule 
coordinated to the magnesium ion has one unperturbed OH giving rise to the band 



137 



3675 




Palygorskite 3400 



Augite 



3000 



3200 



3400 



3600 



3800 



4000 



Wavenumber (cm" ) 
Figure 7.15 DRIFT spectra of the chain and layered silicates in the hydroxyl region 



138 



Topaz 



640 




3570 



Almandite 



3640 3£8v 



Olivine 



3000 



3200 



3400 3600 

Wavenumber (cm" 1 ) 



3800 



4000 



Figure 7.16. DRIFT spectra of orthosilicate minerals in the hydroxyl region. 



139 
at 3619 cm" 1 while the band at 3540 cm 1 arises due to the hydrogen bonding 
between the other hydroxyls on the magnesium ion with neighboring hydroxyls. The 
band at 3400 cm" 1 is due to the presence of molecular water hydrogen bonded to 
the coordinated water molecules on the external surface and the channels inside 
the palygorskite fibers. 
Tremolite and Auaite 

The DRIFT spectra of tremolite shown in Figure 7.10 exhibits several bands 
in the hydroxyl region which are in agreement with those reported by Marel and 
Beutelspacher [Mar76]. On the other hand, a featureless spectra was obtained for 
augite. The presence of (OH) in the amphibole structure is probably responsible for 
the bands in the hydroxyl region while only surface hydroxyls are possible for augite. 
The featureless spectra for augite may then due to a low density of surface 
hydroxyls. 
Almandite. Topaz and Olivine 

The band observed at 3640 cm 1 is attributed to the structural OH for both 
topaz and almandite (see Figure 7.1 1). The bands at 3527 and 3464 cm" 1 for topaz 
are due to the hydrogen bonded hydroxyls. The spectra of olivine provided in 
literature is featureless in the hydroxyl region [Mar76]. Therefore, the band at 3683 
cm" 1 is probably due to an impurity similar to the case of palygorskite coating on 
being present on the dolomite sample reported by Moudgil et al [Mou95b]. 
Correlation between isolated hydroxyls and adsorption 

The different type of hydroxyl groups on the silicate samples along with the 
saturation adsorption density are summarized in Table 7.4. Augite, which did not 



140 



Table 7.4. Type of hydroxyl groups on silicates along with PEO saturation 
adsorption density. 



Material 


Flocculation 


Saturation 

Adsorption 

Density 

(mg/g) 


Type of 
surface 
hydroxyl 


Palygorskite 


Yes 


32.0 


Isolated 
Bonded 


Tremolite 


5.2 


Isolated 
Bonded 


Augite 


3.4 


- 


Almandite 


No 


0.4 


Isolated 


Topaz 


0.6 


Isolated 


Olivine 


0.6 


Isolated 
Bonded 



141 
exhibit isolated hydroxyls, showed a significant amount of PEO adsorption while the 
orthosilicates which all exhibited the isolated hydroxyls showed lower adsorption 
than augite. Thus, similar to the case of oxides, the presence of isolated hydroxyls 
on silicate minerals did not always result in significant PEO adsorption. 
Adsorption Mechanism(s) of PEO on Silicates 

An obvious difference between the flocculating and the non-flocculating 
silicates is the structure of the silicate itself. The chain and layered silicates exhibit 
connectivity of the silicate tetrahedra in one- and two-dimensions respectively and 
show a significantly higher adsorption of PEO than the orthosilicates which have 
silicate tetrahedra isolated from each other through the presence of octahedral 
cavities containing divalent or trivalent atoms. It has been shown earlier that the 
presence of Bronsted acid sites strong enough to bind the ether oxygen of PEO is 
essential for adsorption of PEO. The relation between the structure of the silicates 
and the resultant Bronsted acidity is discussed next. 
Layered silicates 

The surface of the layered silicates consists of broken bonds at the edges, 
hydroxyl groups which are part of the structure, and the exchangeable ions which 
may be coordinated to the former two sites but mainly occur to maintain 
electroneutrality of the tetrahedral or the octahedral layer, when ions of lower 
valence than Si or Al are substituted respectively. 

The Al may substitutes for Si in the silicate layer resulting in enhanced acidity 
of the adjacent SiOH groups. This occurs due to a strong attraction of the bridging 
oxygen between the Al and Si towards Al, which increases the attraction of the 



142 
terminal oxygen on SiOH to Si and weakening of the O-H bond [Hai67]. 
Interestingly, the frequency of the isolated silanol in the mixed oxide system was 
determined to be the same as on silica, 3747 cm V 

In the clays, the isomorphous substitution of Al for Si leading to enhanced 
acidity of the silanol group occurs at the edges only since the layer consists of an 
essentially siioxane surface. However, the replacement of Al for Si in the siloxane 
layer results in a negative charge which is balanced by the exchangeable cation. 
The exchangeable cation by itself is a Lewis acid but due to the presence of a 
hydration layer around the exchangeable ion this acidity is transformed to Bronsted 
acidity. This has been attributed to the submission of the hydration layer to a strong 
polarizing field due to the small radius of curvature of the cation, as compared with 
the quasi-infinite radius of curvature of the anionic siloxane sheet [Fri90]. The high 
strength of the Bronsted acid sites comprising of hydrated exchangeable cations 
and the edge OH groups is well documented through spectroscopic studies 
involving pyridine adsorption and titration with indicator dyes [Rup 87, Fri90, Sch95]. 

The flocculation of montmorillonite, a layered silicate, with PEO has been 
shown to be strongly dependent on the type of exchangeable cation present by 
Scheiner and co-workers [Sch86;Sch87;Bro89]. They suggested that the 
exchangeable ion with its hydration shell constitutes the binding sites for the ether 
oxygen of PEO. Their proposed adsorption mechanism is illustrated in Figure 12. 
However, the flocculation of palygorskite was not significantly affected by the type 
of exchangeable cation [Sch86]. This is expected since the primary source of the 



143 




Figure 7.17. Adsorption of PEO on clays by interaction of the ether oxygen of PEO 
with the hydration shell of the exchangeable ion (Bronsted acid site) (after [Sch86]). 



144 
exchangeable ions on palygorskite has been suggested to be different than on 
montmorillonite [Gri68]. 

The exchangeable cations on the montmorrilonite surface arise as a result 
of isomorphous substitution of Al for Si in the basal planes, as mentioned above. 
The excahngeble cations which are present inside the tactoid are inaccessible to 
the high molecular weight PEO molecules. The increase in the tactoid thickness is 
in the reverse order of the polarizing power of the cations and is accompanied by 
a decrease in the external surface area available for PEO adsorption. Hence, a 
significant lower dosage was required for flocculantion of a Ca-montmorillonite than 
Na-montmorillonite [Sch86]. 

In palygorskite, the edges constitute most of the external surface area in 
contrast to kaolinite and montmorillonite. The edges consist of the broken bonds 
of Si-O-Si type which compensate their residual charge by forming isolated silanol 
groups with a spacing of about 0.5 nm [Rau87]. The edges which form the 
periphery of the palygorskite fibers are also the main source of the exchangeable 
cations. Their absence in the internal structure of the fiber renders the flocculation 
of palygorskite relatively insensitive to the type of exchangeable cation. 

The other source of isolated hydroxyls in palygorskite is the coordinated 
water molecule attached to the magnesium ion. Exposed magnesium ions may play 
a similar role as the exchangeable cations in montmorillonite when the coordinated 
water is replaced. The structural hydroxyls are accessible only at the faces of the 
(100) plane which constitutes a small fraction of the available surface area. This is 
because of the dimensions of the face being less than typically 1/10 th of the fiber 



145 
length. The predominant adsorption of PEO on the edges is indicated by alignment 
of the needles in the floes [Hog85]. 
Chain silicates 

The pyroxene (augite) and amphibole (tremolite) type of silicates do not have 
a significant ion-exchange capacity. Their fibrous morphology, similar to 
palygorskite, presents the Si-O-Si broken bonds at the surface which can satisfy the 
valence by formation of SiOH type of groups. This continuity of SiOH type of sites 
is lost in the orthosilicates where the silicate tetrahedra are isolated. The influence 
of the polymerization of Si04 tetrahedra on the acidity of these SiOH groups needs 
to be investigated. The adsorption of PEO on tremolite and augite indicates that 
SiOH groups are more acidic than on the orthosilicates, and are close in acidity to 
the SiOH on layered silicates. 

The surface characterization of the silicate samples revealed that strong 
Bronsted sites capable of interaction with the ether oxygen of PEO were present 
only on layered and chain silicates such as kaolinite, palygorskite, tremolite and 
augite. The presence of such strong Bronsted sites seems to be related to the 
connectivity of the silicate tetrahedra since the orthosilicates did not exhibit 
significant PEO adsorption. Thus the adsorption mechanism for PEO determined 
on silica was shown to be the same for mixed oxides containing silica. 



CHAPTER 8 

CONCLUSIONS AND FUTURE WORK 

Summary 

The specificity of hydrogen bonding of the ether oxygen of PEO with the 
surface hydroxyls (Bronsted sites) has been established in this study. The acidity 
of the Bronsted sites is the parameter that determines the interaction with the ether 
oxygen of PEO, a Lewis base. The strength of the Bronsted acid sites on oxide 
samples was characterized by their point of zero charge and vibrational 
characteristics of adsorbed pyridine, a weak base. It was determined that highly 
acidic oxides of the type M0 3 , Iv^Q and MQ such as Mop , 2 V 5 , and §iO 
strongly adsorb and flocculate with PEO. Ti0 2 , an MQ type of oxide, did not 
adsorb PEO due to weaker acidity of the Bronsted sites than the silanols. 

It was shown that the common feature of the previously identified PEO 
binding sites, the isolated silanols on silica and exchangeable ions on clays, is the 
acid strength of these sites. The Bronsted acid sites on clays also consists of the 
exchangeable ions besides the isolated hydroxyls. The concept of the strength of 
the Bronsted acid sites governing the interaction with the ether oxygen of PEO thus 
explained all the previous and present observations made regarding the adsorption 
sites of PEO on oxide and silicate surfaces. 



146 



147 
This study also revealed that the presence of isolated surface hydroxyls, a 
criterion extrapolated from adsorption studies on silica/non-ionic polymer system, 
is not sufficient to explain adsorption of PEO on oxide surfaces. The acidity of 
isolated silanols arises from the nature of the silica surface. The oxides bearing 
isolated hydroxyls on the surface, e.g., alumina, titania, magnesia and hematite did 
not exhibit measurable PEO adsorption. On the other hand crystalline silica, Mo0 3 
and V 2 5 which did not exhibit the isolated hydroxyls in their spectra showed 
significant PEO adsorption. Similarly, for the silicate samples examined no 
correlation between the presence of isolated surface hydroxyls and PEO adsorption 
was observed. 

The specificity of hydrogen bonding of isolated silanols with polar organic 
molecules such as diethyl ether, the monomer unit of PEO, has been established 
for the case of adsorption from the vapor state [Kis65]. In the present study it was 
determined that the specificity of hydrogen bonding interaction between the isolated 
silanols and the ether oxygen of PEO persists in the presence of an aqueous 
medium. The presence of water molecules, however, precludes this interaction to 
be observed by infrared spectroscopy. The observation of only the isolated silanols 
in the DRIFT spectra of silica pretreated at 1 100°C along with significant adsorption 
of PEO on this sample indicated that the isolated silanols are indeed the principal 
binding sites for the ether oxygen of PEO. 

The isolated silanols were shown to be the stable surface sites for PEO 
adsorption with respect to pH. The adsorption of PEO on silica A, which exhibited 
isolated silanols, was insensitive to pH in the range 3.0 to 9.5. On the other hand, 
PEO adsorption was significantly reduced on silica B, which did not exhibit the 



148 
isolated silanols, when the pH was changed from 3.0 to 9.5. The desorption of high 
molecular weight PEO molecules from a fused silica plate was shown to occur by 
AFM studies when the pH was changed in-situ from 3.0 to 9.5. 

The adsorption of PEO on silica B and silica plate surface near the isoelectric 
point of pH 2.0, however, indicated that silanols other than the isolated type are 
capable of hydrogen bonding with the ether oxygen of PEO. It is suggested that 
their low concentration and higher acidity compared to the isolated silanols results 
in significant ionization of these sites with increasing pH and thus a concomitant 
loss of PEO adsorption sites. The electrokinetic and PEO adsorption studies on 
silica B support this hypothesis. Further compelling evidence, showing the loss of 
PEO binding sites accompanied by an increasingly negatively charged surface on 
a silica plate, was illustrated in the AFM studies. The desorption of PEO molecules 
upon changing the pH from 3.0 to 9.5 was corroborated by the force/distance 
profiles which changed from that for a pure steric repulsion to completely 
electrostatic repulsion for the same pH change. The AFM investigations thus 
showed that previous hypotheses regarding the observed decrease in adsorption 
of non-ionic polymers like PEO with increasing pH such as 1) a negatively charged 
surface repels PEO molecules and 2) the hydrated counter ions prevent the 
approach of PEO to the silica surface are not correct. 

The surface acidity of the oxides and silicates may also arise from the 
presence of Lewis acid sites. Among the oxides investigated in the present study 
the strongest Lewis acid sites have been demonstrated on the alumina surface 
[Dav90]. The heat pretreatment of alumina, which resulted in a partially 



149 
dehydroxylated surface and an increased concentration of Lewis acid sites, was 
found not to affect the PEO adsorption significantly. It appears, therefore, that the 
Lewis acid sites do not play a major role in PEO adsorption, a conclusion also 
supported by the investigations of van der Beek et al [van91]. 

The dehydroxylation process also mitigates the accessibility of the Bronsted 
acid sites on the alumina surface to the PEO molecules. However, the PEO 
adsorption was not affected significantly indicating that the lack of accessibility to 
the surface sites is not the reason for negligible PEO adsorption on alumina. It is 
the presence of Bronsted sites of weaker acidity than the silanols on the alumina 
surface which results in an insignificant adsorption of PEO on alumina. 

The Bronsted acid sites on silicates capable of binding the ether oxygen of 
PEO were shown to be related to the connectivity of the Si0 4 tetrahedra in chains 
or layers. Thus layered silicates such as kaolinite and palygorskite and chain 
silicates such as pyroxenes (augite) and amphiboles (tremolite) were shown to 
adsorb PEO while the orthosilicates such as olivine, almandite and topaz did not 
exhibit PEO adsorption. 

The other suggested PEO adsorption mechanisms in the literature such as 
the presence of adsorbed surface ions, which in solution state are capable of 
complexing with the ether oxygen, and positively charged surface sites do not play 
a major role in the adsorption process. The presence of sodium ions on the 
alumina B surface and precipitation of Mo ions on alumina and titania did not affect 
the adsorption of PEO on these surfaces. Also, predominance of positively charged 
surface sites on alumina or hematite at pH 3.0 did not result in significant PEO 



150 
adsorption. On the other hand, the high negative charge on the surface of silica A, 
Mo0 3 and V 2 5 did not preclude PEO adsorption. 

Sug gestions for Future Work 

The concept of the specificity of hydrogen bonding of PEO with the surface 
hydroxyls need to be extended to other non-ionic polymers such as PVA and 
polyacrylamide (PAM). The behavior of PVA/silica system has been shown to be 
similar to the PEO/silica system[Tad78, Kha88]. PAM is known to be a stronger 
flocculant than both PEO and PVA [Sch87]. This indicates that the surface/segment 
interactions for PAM are stronger than PEO and PVA. The quantification of the 
substrate-polymer-solvent interactions needs to be developed to predict the 
adsorption behavior at the solid/solution interface. 

The characterization of adsorption sites on oxides was accomplished through 
electrokinetic studies and infrared spectroscopy of adsorbed pyridine. However, 
determination of the concentration and strength distribution of surface sites will be 
of value in identifying methods to create the appropriate sites for polymer 
adsorption. For example, Stober silica and crystalline silica exhibited different PEO 
adsorption with respect to pH which was attributed to the presence of isolated 
silanols only on Stober silica. The elucidation of the surface structure of the two 
silica samples by molecular modeling and experimental techniques is necessary for 
developing of pretreatments that can result in the formation of isolated silanols on 
crystalline silica. 

In addition to elucidation of the surface structure and the distribution and 
reactivity of surface sites, molecular modeling of the polymer molecules will offer 



151 
valuable insight into the adsorption process. The combined knowledge of the 
molecular architecture of the surface and reactivity of the different possible 
adsorption sites along with the chemistry of the polymer molecules will be useful in 
synthesis/identification of an appropriate polymer for a given particulate system. 

The potential of AFM to investigate the microstructure at the solid/solution 
interface in the system of intrest has been demonstrated in this study. The cause 
of the high saturation adsorption density of PEO determined on the layered and 
chain silicates was investigated through AFM studies. Although no explanation at 
present seems to exist, information on PEO conformation on a flat surface was 
obtained. The polymer appeared to adsorb with a flatter conformation on tremolite 
than augite as indicated by the AFM image analysis. A definite comparison between 
the parking area of the adsorbed polymer molecules and their size in solution could 
not be made due to the polydispersity of the polymer samples. It is suggested that 
further detailed investigations with monodisperse polymers can reveal useful 
quantitative information on polymer conformation at the solid/liquid interface. A 
comparison of the experimental data so obtained with the theoretical predictions will 
lead to a better understanding of the adsorbed polymer layer conformation and 
subsequent control of the stability of suspensions. 



REFERENCES 



[Ana87] 

[And 65a] 
[And82] 

[And65b] 

[Arm69] 

[Att87] 



[Bai76] 
[Bas61] 

[Beh93a] 

[Beh93b] 



Ananthapadmanabhan, K.P. and Goddard, E.D.,"Aqueous Bi-Phase 
Formation of Polyethylene Oxide-Inorganic Salt Systems," Langmuir, 
3(1), pp. 25-31,(1987). 

Anderson J.H., Jr., "Calorimetric vs. Infrared Measures of Adsorption 
Bond Strengths on Silica," Surface Science, 3, pp. 290-291, (1965). 

Anderson, M.A., and Rubin, A.J., Adsorption of Inorganics at 
Snliri/l ig uid Interface . Ann Arbor Science Publishers, Ann Arbor, Ml, 
USA, (1981). 

Anderson, P.J., Horlock, R.F., and Oliver, J.F., "Interaction of Water 
with the Magnesium Oxide Surface/Transactions of the Faraday 
Society, 61, pp. 2754-2762, (1965). 

Armistead, C.G., Tyler, A.J., Hambleton, F.H., Mitchell, S.A., and 
Hockey, J. A., "The Surface Hydroxylation of Silica," Journal of 
Physical Chemistry, 73, pp. 3947-3953, (1969). 

Attia, Y.A., "Design of Selective Polymers for Biological and Colloid 
Separation by Selective Flocculation," in Flocculation in Biotechnology 
and Separation Systems, Ed., Attia, Y.A., Elsevier Science 
Publishers, pp. 297-308, (1987). 

Bailey Jr., F.E., and Koleske, J.V., Polyethylene oxide), Academic 
Press, New York, (1976). 

Basila, M.R., "Hydrogen Bonding Interaction Between Adsorbate 
Molecules and Surface Hydroxyl Groups on Silica," Journal of 
Chemical Physics, 35(4), pp. 1151-1158, (1961). 

Behl, S., "Controlled Polymer Adsorption for Enhanced Selectivity in 
flocculation," Ph.D. Thesis, University of Florida, Gainesville, FL, USA, 
(1993). 

Behl, S., Moudgil, B.M., and Prakash, T.S., "Control of Active Sites in 
Selective Flocculation -Part I: A Mathematical Model", Journal of 
Colloid and Interface Science, 161, pp. 414-421, (1993). 



152 



153 



[Beh93c] Behl, S. and Moudgil, B.M.,"Role of Active Sites in Flocculation - 
Concept of Equivalent Active Sites", Journal of Colloid and Interface 
Science, 161, pp. 437-442, (1993). 

[Beh93d] Behl, S., and Moudgil, B.M., "Mechanisms of Polyethylene Oxide 
Interaction with Apatite and Dolomite", Journal of Colloid and 
Interface Science, 161, pp. 443-449, 1993. 

[Ber94] Berhouet, S., and Toulhoat, H., "Nature of Acidic Sites of the Lewis 

Type at the surface of Mica and related Aluminosilicate Materials," 
Langmuir, 10, pp. 1832-1836, (1994). 

[Bla90] Blaakmeer, J., "The Adsorption of Weak Polyelectrolytes and 

Polyampholytes," Ph.D. Thesis, Agricultural University, Wageningen, 
The Netherlands, (1990). 

[Bri90] Brinker, C.J., and Scherer, G W Sol-Gel Science: The Physics and 

Chemistry of Sol-Gel Processing . Academic Press, Boston, MA, USA, 
(1990). 

[Bro89] Brown, P.M., Stanley, DA, and Scheiner, B.J./'An explanation of 

flocculation using Lewis acid-base theory", Minerals and Metallurgical 
Processing, pp. 196-200, (1989). 

[Car94] Carniti, P., Gervasini, A., Auroux, A., "Energy Distribution of Surface 

Acid Sites of Metal Oxides," Journal of Catalysis, 150, pp. 274-283, 
(1994). 

[Cha86] Chanchani, R., "Selective Flotation of Dolomite from Apatite Using 

Sodium Oleate as the Collector,"Ph.D. Thesis, University of Florida, 
Gainesville, FL, USA, (1986). 

[Cha63] Chapman, ID. and Hair, M.L/The Role of the Surface Hydroxyl 

Groups in Catalytic Cracking," Journal of Catalysis, 2, pp. 145-148, 
(1963). 

[Che85] Cheng, Y.C./The Effect of Surface Hydration on the Adsorption and 

Flocculation of a Model Silica Suspension Using Polyethylene Oxide," 
M.S. Thesis, University of Florida, Gainesville, Florida, USA, (1985). 

[Che59] Chessick, J.J., and Zettlemoyer, A.C. /'Heats of Immersion of Solids 
in Liquids/Advances in Catalysis, 11, pp. 263-286, (1959). 



154 



[Chr93] Chronister, C.W., and Drago, R.S. /'Determination of Hydrogen- 

Bonding Acid Sites on Silica using the Cal-Ad Method," Journal of the 
American Chemical Society, 115, pp. 4793-4798, (1993). 

[Coh84] Cohen-Stuart, MA, Waajen, F.H.W.H., Cosgrove,T., Vincent, B., and 

Crowley, T. L. , "Hydrodynamic Thickness of adsorbed Polymer Layers," 
Macromolecules, 17(9), pp. 1825-1830, (1984). 

[Cos84] Cosgrove, T., Vincent, B., Crowley, T.L., and Cohen Stuart, M.A., 

"Segment Density Profiles of Adsorbed Polymers," in Polymer 
Adsorption and Dispersion Stability . Eds. Goddard, E.D., and Vincent, 
B., American Chemical Society, Washington, D.C., pp.147-159, 
(1984). 

[Cos90] Cosgrove, T., "Volume Fraction Profiles of Adsorbed Polymers," 

Journal of Chemical Society Faraday Transactions, 86(9), pp. 1323- 
1332, (1990). 

[Cow83] Cowell, C. And Vincent, B., "Flocculation Kinetics and Equilibria in 
Sterically Stabilized Dispersions," Journal of Colloid and Interface 
Science,95(2), pp.573-582, (1983). 

[Cul94] Cullen, D.C., and Lowe, C.R., "AFM Studies of Protein Adsorption", 

Journal of Colloid and Interface Science, 166, pp.102-108, (1994). 

[Dav90] Davydov, AA." lnfrared Spectroscopy of Adsorbed Species on the 

Surface of Transition Metal Oxides." Ed. Rochester, C.H., John Wiley 
&Sons, New York, (1990). 

[deG81] de Gennes, P. G. ."Polymer Solutions Near an Interface I: Adsorption 

and Depletion Layers," Macromolecules, 14, pp. 1637-1644, (1981). 

[deG82] de Gennes, P.G., "Polymers at an Interface II: Interaction between 

Two Plates Carrying Adsorbed Polymer Layers," Macromolecules, 15, 
pp.492-500, (1982). 

[deG87] de Gennes, P. G., "Polymers at an Interface; a Simplified View," 

Advances in Colloid and Interface Science, 27, pp. 189-209, (1987). 

[Dow92] Dow, J.H., "Dispersion of Ceramic Powders in Ceramic Melts, "PhD 
Thesis, University of Florida, Gainesville, Florida, USA, (1992). 

[Eva73] Evans, R., and Napper, D.H., "Flocculation of Latices by Low 

Molecular Weight Polymers," Nature, 246, pp. 34-35, (1973). 



155 

[Far64] Farmer, V.C., and Russel, J.D., "The Infra-red Spectra of Layer 

Silicates," Spectrochimica Acta, 20, pp. 1149-1173, (1964). 

[Feh89] Feher, F.J., Newman, D.A., and Walzer, J.F., "Silsesquioxanes as 

Models for Silica Surfaces," Journal of American Chemical Society, 

111, pp. 1741-1748, (1989). 

[Feh90] Feher, F.J., and Newman, D.A., "Enhanced Silylation Reactivity of a 

Model for Silica Surfaces," Journal of American Chemical Society, 

112, pp. 1931-1936, (1990). 

[Fle83] Fleer, G.J., and Lyklema, J., "Adsorption of Polymers," in Adsorption 

from Solution at Solid/Liquid Interface . Eds. Parfitt, G.D., and 
Rochester, C.H., Academic Press, London, pp. 153-222, (1983). 

[Fri62] Fripiat, J.J., and Uytterhoeven, J.,"Hydroxyl Content in Silica Gel 

Aerosil," Journal of Physical Chemistry, 66, pp. 800-805, (1962). 

[Fri90] Fripiat, J.J., "Surface Activities of Clays," in Spectroscopic 

Characterization of Minerals and Their Surfaces , Coyne, L.M., 
McKeever, W.S., and Blake, D.F., eds., American Chemical Society, 
Washington, DC, USA, pp.360-377, (1990). 

[Gra93] Graabe, A., and Horn, R.G., "Double-Layer and Hydration Forces 

Measured between Silica Sheets Subjected to Various Surface 
Treatments," Journal of Colloid and Interface Science, 157, pp. 375- 
383, (1993). 

[Gre72a] Greenland, D.J., "Interactions between organic polymers and 
inorganic soil particles," Mededelingen Fakulteit Landbou- 
wwetenschappen Rijksuniversiteit Gent, 37, pp. 897-91 4 (1972). 

[Gre72b] Greenland, D.J., "Interactions between organic polymers and 
inorganic soil particles," Mededelingen Fakulteit Landbou- 
wwetenschappen Rijksuniversiteit Gent, 37, pp.91 5-922, (1972). 

[Gri68] Grim, R.E., Clay Mineralogy . McGraw Hill Inc., New York, NY, USA 

(1968). 

[Hai67] Hair, M.L., Infrared Spectroscopy in Surface Chemistry . Marcel 

Dekker, New York, NY, USA, (1967). 

[Hea65] Healy, T.W., and Fuerstenau, D.W., "The Oxide-Water Interface- 

Intrerelation of the Zero Point of Charge and the Heat of Immersion," 
Journal of Colloid Science, 20, pp. 376-386, (1965). 



156 



[Hog85] Hoghooghi, B., "Flocculation of Palygorskite Suspensions and the 

Orientation of Particle in Floes, "M.S. Thesis, University of Florida, 
Gainesville, FL, USA, (1985). 

[How67] Howard, G.J., and McConell, P., "Adsorption of Polymers at the Solid 
Solution Interface," Journal of Physical Chemistry, 71(9), pp. 2974- 
2980, (1967). 

[Ile71] Her, R.K., The Chemistry of Silica . John Wiley and Sons, New York, 

NY, USA, (1971). 

[Ile75] Her, R.K.,"The Effect of Surface Aluminosilicate Ions on the Properties 

of Colloidal Silica," Journal of Colloid and Interface Science, 55(1), pp. 
25-34, (1975). 

[Iwa82] Iwasaki, I., and Lipp, R.J., in Interfacial Phenomena in Mineral 

Processing , eds. Yarar, B., and Spottiswood, D.J., Engineering 
Foundation, New York, NY, pp. 321, (1982). 

[Kha88] Khadilkar.C.S., "The Effect of Adsorbed Polyvinyl alcohol) on the 

Properties of Model Silica Suspensions," Ph.D. Thesis, University of 
Florida, Gainesville, Florida, USA, (1988). 

[Kho84] Khosla, N.K., Bhagat, R.P., Gandhi, K.S., and Biswas,A.K. 

"Calorimetric and Other Interaction Studies on Mineral-Starch 
Adsorption System," Colloids and Surfaces, 8, pp. 321-336, (1984). 

[Kin76] Kingery, W.D., Introduction to Ceramics . Second Edition, John Wiley 

& Sons, Inc., New York, NY, USA, (1976). 

[Kis65] Kiselev, A.V., "Calorimetric vs. Infrared Measurements of Adsorption 

Bond Strengths of Molecules of Different Electronic Structure on 
Hydroxylated Silica Surface," Surface Science, 3, pp. 292-293, 
(1965). 

[Kje81] Kjellander, R., and Florin, E., "Structure of Polyethylene Oxide in 

Water," Journal of Chemical Society, Faraday Transaction I, 77, pp. 
2053,(1981). 

[Kle84] Klein, J., "Forces Between Two Adsorbed Poly(ethylene oxide) Layers 

in a Good Aqueous Solvent in the Range 0-150 nm," 
Macromolecules, 17, pp. 1041-1048, (1984). 



157 



[Kok90] Koksal, E., Ramachandran, R., Somasundaran, P., and Maltesh, C, 

"Flocculation of Oxides using Poly(ethylene oxide)," Powder 
Technology, 62, pp. 253-259, (1990). 

[Kun89] Kung, H.H., Transition Metal Oxides: Surface Chemistry and 

Catalysis , Elsevier, Amsterdam, The Netherlands, (1989). 

[Luc90] Luckham, P.F., and Klein, J., "Force between Mica Surfaces Bearing 

Adsorbed Homopolymers in Good Solvents," Journal of Chemical 
Society Faraday Transactions, 86(9), pp. 1363-1368, (1990). 

[Mor76] Morterra, C, Ghiotti, G., Garrone, E., and Bocuzzi, F., "Infrared 

Spectroscopic Characterization of the a-AI 2 3 Surface," Journal of 
Chemical Society, Faraday Transactions 1 , 72, pp.2722-2734, (1976). 

[Mor68] Mortland, M.M., and Raman, K.V., "Surface Acidity of Smectites in 

Relation to Hydration, Exchangeable Cation and Structure," Clays and 
Clay Minerals, 16, pp. 393-398, (1968). 

[Mou91] Moudgil, B.M., and Behl, S., "A Model of the Selective Flocculation 

Process, "Journal of Colloid and Interface Science, 146(1), pp. 1-8, 
(1991). 

[Mou92] Moudgil, B.M., Behl, S., and Kulkarni, N. ."Measurement of Heat 

Adsorption of Polyethylene Oxide on Dolomite, Silica and Alumina by 
Microcalorimetry," Journal Colloid Interface Science, 148(2), pp. 337- 
342, (1992). 

[Mou95a] Moudgil, B.M., Mathur, S., and Behl, S., "Dolomite-Dolomite 
Separation by Selective Flocculation Technique," Minerals and 
Metallurgical Processing, 12 (1), pp. 24-27, (1995). 

[Mou95b] Moudgil, B.M., Mathur, S. and Behl, S, "Flocculation Behavior of 
Dolomite With Poly(ethylene-oxide)," Minerals and Metallurgical 
Processing, 12(4), pp. 219-224, (1995). 

[Oka88] Okamoto, Y., and lmanaka,T., "Interaction Chemistry between 

Molybdena and Alumina: Infrared Studies of Surface Hydroxyl groups 
and Adsorbed Carbondioxide on Aluminas Modified with Molybdate, 
Sulfate, or Fluorine Atoms," Journal of Physical Chemistry, 92, pp. 
7102-7112, (1988). 

[Par65] Parks, G.A., "Isoelectric Points of Solid Oxides, Solid Hydroxides and 

Aqueous Hydroxo Complex Systems," Chemical Review, 65, pp. 177- 
197, (1965). 



158 



[Par63] Parry, E.P., "An Infrared Study of Pyridine Adsorbed on Acidic Solid- 

Characterization of Surface Acidity", Journal of Catalysis, 2, pp. 371- 
379, (1963). 

[Pau63] Pauling, L. Nature of the Chemical Bond . Cornell University Press, 

Ithaca, New York, USA, (1963). 

[Pel90] Pelssers, E.G.M., Cohen Sturat, M.A., and Fleer, G.J., "Kinetics of 

Bridging Flocculation-Role of Relaxations in the Polymer 
Layer,"Journal of Chemical Society, Faraday Transactions, 86(9), pp. 
1355-1361, (1990). 

[Pra91] Pradip, "On the Design of Selective Reagents for Mineral Processing 

Applications," Metals, Materials & Processes, 3 (1), pp. 15-36, (1991). 

[Pra96] Prakash, T.S., "Separation of Salts Using Polymer Bi-Phase System," 

M.S. Thesis, University of Florida, Gainesville, Florida, USA, (1996). 

[Pri71] Primet, M., Pichat, P., and Mathieu, M.V., "Infrared Study of the 

Surface of Titanium Dioxides, I. Hydroxyl Groups," Journal of Physical 
Chemistry, 75(9), pp. 1216-1220, (1971). 

[Ram89] Ramachandran, R., and Somasundaran, P., "Polymer Surfactant 
Interactions in bulk and at the Solid/Liquid Interface", in Flocculation 
and Dewatering . eds. Moudgil, B.M., and Scheiner, B.J., Engineering 
Foundation, New York, United Engineering Trustees, Inc., New York, 
NY, (1989). 

[Rau87] Raussell-Colom, J. A., and Serratosa, J.M., "Reactions of Clays with 

Organic Substances," in Chemistry of Clavs and Clav Minerals , ed. 
Newman, A.C.D., John Wiley & Sons, New York, NY, USA, pp. 371- 
422, (1987). 

[Ree88] Reed, J.S., Principles of Ceramic Processing . John Wiley & Sons Inc., 

New York, NY, USA, (1988). 

[Rob64] Robinson, M., Pask, J.A., and Fuerstenau, D.W., "Surface Charge of 

Alumina and Magnesia in Aqueous Media, "Journal of the American 
Ceramic Society, 47(10), pp. 516-520, (1964). 

[Roc76] Rochester, C.H., "Infrared Spectroscopic Studies of Powder Surfaces 

and Surface-Adsorbate Interactions Including the Solid/Liquid 
Interface. A Review," Powder Technology, 13, pp. 157-176, (1976). 



159 



[Rub76] Rubio, J. and J. A. Kitchener/The Mechanism of Adsorption of 

Poly(Ethylene Oxide) Flocculant on Silica," Journal of Colloid and 
Interface Science, 57(1), pp. 132-141, (1976). 

[Rup87] Rupert, J. P., Granquist, W.T., and Pinnavaia, T.J., "Catalytic 

Properties of Clay Minerals," in Chemistry of Clays and Clay Minerals , 
ed. Newman, A.C.D., John Wiley & Sons, New York, NY, USA, pp. 
275-314, (1987). 

[Sch82] Scheutjens, J.M.H.M., and Fleer, G.J., "Adsorption of Interacting 

Oligomers and Polymers at an Interface," Advances in Colloid and 
Interface Science, 16, pp. 341-359, (1982). 

[Sch86] Scheiner, B.J. and Stanley, D.A.,"Flocculation of fine particles with 

polyethylene oxide: a proposed mechanism", Transactions of the 
Society of Mining Engineers of AIME, 280, pp. 21 1 5-21 1 7, (1 986). 

[Sch87] Scheiner, B.J. and Wilemon, G.M., "Applied Flocculation Efficiency: A 

Comparison of Polyethylene Oxide and Polyacrylamide," in 
Flocculation in Biotechnology and Separation Systems . Ed. Attia, 
Y.A., Elsevier Science Publishers, Amsterdam, pp. 175-186, (1987). 

[Sch95] Schoonheydt, R.A., "Clay Mineral Surfaces," in Mineral Surfaces , eds. 

Vaughan, D.J., and Pattrick, R.A.D., Chapman & Hall, London, 
England, pp. 303-332, (1995). 

[Ser77] Serna, C, VanScoyoc, G.E., and Ahlrichs, J.L., "Hydroxyl Groups and 

Water in Palygorskite," American Mineralogist, 62, pp. 784-792, 
(1977). 

[Sha86] Shah, B.D., "Selectivity in Mixed Mineral Flocculation: Apatite- 

Dolomite System," M.S. Thesis, University of Florida, Gainesville, FL, 
USA, (1986). 

[Spe92] Sperling, L.H., Introduction to Physical Polymer Science . Second 

edition, John Wiley & Sons, Inc., NY, USA, (1992). 

[Tad78] Tadros, Th. F., "Adsorption of Polyvinyl alcohol) on Silica at pH 

Values and its Effect on the Flocculation of the Dispersion," Journal 
of Colloid and Interface Science, 64(1), pp. 36-47, (1978). 

[Tro83] Tronel-Peyroz, E., Raous, H., and Schuhmann, D., "A Study of the 

Interfacial behavior of Polyoxyethylene at a Mercury-Aqueous 
Solution Interface: Compact and Diffuse Layers," Journal of Colloid 
and Interface Science, 92(1), pp. 136-153, (1983). 



160 



[Tsy72] Tsyganenko, A.A., and Filimonov, V.N., "Infrared Spectra of Surface 

Hydroxyls Groups and Crystalline Structures of Oxides," Spectroscopy 
Letters, 5(12), pp. 477-487, (1972). 

[van 91] van der Beek, G.P., Cohen Stuart, M.A., Fleer, G.J., and Hofman, 
J.E., "Segmental Adsorption Energies for Polymers on Silica and 
Alumina," Macromolecules, 24, pp. 6600-6611, (1991). 

[van 76] van der Marel, H.W., and Beutelspacher, H., "Atlas of Infrared 
Spectroscopy of Clay Minerals and Their Admixtures." Elsevier 
Scientific Publishing Company, Amsterdam, The Netherlands, pp. 
322-329, (1976). 

[Vas86] Vassilev, K.G., Dimov, D.K., Stamenova, R.T., Boeva, R.S., and 

Tsvetanov, Ch. B., "Complex Forming Properties of CrossLinked 
Poly(ethylene oxide). I. Interaction of Linear and CrossLinked 
Polyethylene oxide) with Molybdenum VI Salts," Journal of Polymer 
Science:Part A: Polymer Chemistry, 24, pp. 3541-3554, (1986). 

[Wal96] Walker, W.J., Reed, J.S., Verma, S., and Zirk, W., "Adsorption 

Behavior of Polyethylene Glycol", Paper No. B-101-96, presented at 
the 98th Annual Meeting of the American Ceramic Society, 
Indianapolis, IN, (1996). 

[Wei95] Weissenborn, P.K., Warren, L.J., and Dunn, J.G., "Selective 

Flocculation of Ultrafine Iron Ore. 1. Mechansim of adsorption of 
starch onto hematite," Colloids and Surfaces, 99, pp. 11-27, (1995). 

[You58] Young, G.J., "Interaction of Water Vapor With Silica Surfaces," 

Journal of Colloid Science, 13, pp. 67-85, (1958). 



BIOGRAPHICAL SKETCH 
Sharad Mathur was born in Kotputli, India, on December 14, 1966. After 
completing his primary education at the St. Paul's School, Kota, he entered the 
bachelor's program in ceramic engineering at the Institute of Technology, Benaras 
Hindu University, Varanasi, India. He graduated in May 1988 and proceeded to 
Canada in September 1988 to pursue further studies at McMaster University, 
Hamilton. After obtaining his M.Eng. degree in materials science and engineering 
in August 1991, he took a four month hiatus in India. He joined the doctoral 
research program in the Department of Materials Science and Engineering at the 
University of Florida in January, 1992. 



161 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope 
and quality, as a dissertation for the degree of Doctor of Philosophy. 




idail. Chairman vj 



Brij MCMoudgil, Chairman 
Professor of Materials 

Science and Engineering 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate?, in scope 
and quality, as a dissertation for the degree of Doctqpof Philosophy.] 





E. Dow Whitney 
Professor of Materij 

Science and Engineering 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope 
and quality, as a dissertation for the degree of Doctor of Philosophy. 



Christopher D. Batich 
Professor of Materials 

Science and Engineering 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope 
and quality, as a dissertation for the degree of Doctor of Philosophy. 

Hassan El-Shall 
Engineer of Materials Science 
and Engineering 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope 
and quality, as a dissertation for the degree of Doctor of Philosophy. 




Rajiv K. Singh 
Associate Professor of 

Materials Science and 

Engineering 



I certify that I have read this study and that in my opinion it conforms to 
acceptable standards of scholarly presentation and is fully adequate, in scope 
and quality, as a dissertation for the degree of Doctor of Philosophy. 



Dinesh O. Shah 
Professor of Chemical 
Engineering 



This dissertation was submitted to the Graduate Faculty of the College of 
Engineering and to the Graduate School and was accepted as partial fulfillment 
of the requirements for the degree of Doctor of Philosophy. 



December 1996 



/— Winfred M. F 



Winfred M. Phillips 

Dean, College of Engineering 



Karen A. Holbrook 
Dean, Graduate School 






LD 

1780 

1994 

,M3Z 



UNIVERSITY OF FLORIDA 



3 1262 08555 0662